Ram934/mpnet-base-all-nli-triplet

BGE base Financial Matryoshka

This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5 on the json dataset. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: BAAI/bge-base-en-v1.5
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity
• Training Dataset:
  • json
• Language: en
• License: apache-2.0

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
  (2): Normalize()
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("Ram934/mpnet-base-all-nli-triplet")
# Run inference
sentences = [
    "[' # ESTIMATES: Accounting estimates\\n  ##  7 Obtain an understanding of the estimation SCOTs\\n    Obtaining an understanding of the estimation SCOTs for lower risk estimates is generally less extensive and time-consuming as lower risk estimates may not require significant judgments and the estimation SCOT is likely to be less complex. When our audit strategy for an accounting estimate is a substantive only approach, we may obtain our understanding of the process used to develop the estimate as we perform our substantive procedures over the estimate. In other situations, when an estimation process is more formal, more extensively supported by IT applications, and carried out throughout the audit period, such as for more complex estimates in the financial services industry, we may perform these procedures earlier in the audit and separate to our substantive procedures. We may also perform our procedures to identify and understand controls that respond to significant risks at the same time as our substantive procedures (refer to  ESTIMATES 10.1). However, we consider whether these controls are likely to be application and ITDM controls for which we will need to understand and evaluate ITGCs relevant to these controls (i.e., those that meet the criteria in  SCOTS 6.1). In these situations, we may perform these procedures earlier in the audit and separate to our substantive procedures. We recognize that the identification and assessment of risks of material misstatement is an iterative process. Therefore, as we obtain an understanding of the process, we may identify further risks of material misstatement (i.e., what can go wrongs) that require additional audit responses (e.g., further tests of controls and/or substantive procedures). In some situations, our understanding of the process may result in a reassessment of our preliminary assessment of inherent risk as required in  ESTIMATES 4 and therefore our categorization of the estimate as higher or lower risk. When understanding the estimation SCOT, we consider whether there are policies that we would expect to be in place at the entity but are absent, or that are not designed to allow management or employees, in the normal course of performing their assigned functions, to prevent or detect and correct misstatements on a timely basis (i.e., the policies established by management may not result in the reporting of transactions in the financial statements in accordance with the applicable financial reporting framework). A lack of appropriate policies may be an indicator of a control deficiency, and we consider whether a risk of material misstatement (i.e., WCGW) exists. Based on our understanding of the critical path of all SCOTs, including our understanding of estimation SCOTs, the significant disclosure processes and the FSCP sub-processes, we evaluate whether together they appropriately support the preparation of the entity’s financial statements in accordance with the applicable financial reporting framework. We also determine whether we have identified one or more control deficiencies. Refer to  FSCP 6.', ' # SCOTS: Significant classes of transactions [effective for audits of periods ending before 15 December 2022]\\n  ##  7 Confirm our understanding of routine and non-routine SCOTs\\n   We confirm our understanding of routine and non-routine SCOTS to:   Confirm the SCOT is correctly documented  Test whether our documentation reflects what actually occurs  Assess whether we have identified relevant WCGWs For estimation SCOTs, refer to  ESTIMATES 11. We confirm our understanding and documentation of a SCOT through walkthrough procedures, which include a combination of inquiry, observation, inspection, and tracing a transaction through the critical path of the SCOT (i.e., through initiation, recording, processing and reporting). ', ' # ESTIMATES: Accounting estimates\\n  ##  6 Identify estimation SCOTs\\n  <REQUIREMENT> We identify a separate estimation SCOT for each accounting estimate that is susceptible to material misstatement. When an accounting estimate is a significant disclosure, we identify a significant disclosure process.  ISA 540.13, AICPA AU-C 540.12</REQUIREMENT>  We recognize that estimation processes are often less formal and are performed at or near the period end. Estimation processes generally do not follow a flow of transactions but contain a series of actions management follows to determine the estimate.   When an estimate in a significant account has related disclosures, we either include the process to develop the related disclosures in the estimation SCOT or we identify a separate significant disclosure process for the related disclosures.', '# ESTIMATES: Accounting estimates\\n  ##  11 Confirm our understanding of the estimation SCOTs\\n<not_applicable_for_pcaob>  Controls reliance strategy We determine whether our understanding of the estimation SCOT is appropriate based on our substantive procedures. <REQUIREMENT> When we take a controls reliance strategy, we confirm our understanding of, and evaluate, the design of the relevant controls in estimation SCOTs every audit period.  ISA 315.26, AICPA AU-C 315.26-30</REQUIREMENT> Estimation SCOTs typically do not contain transaction flows but consist of a series of actions within a process to determine the estimates. Therefore, instead of tracing one transaction along the critical path, we confirm our understanding of the design of the relevant controls in the estimation SCOT by:    Inquiring of the process owners about the series of actions within the process to determine whether our understanding and documentation of relevant controls is accurate  Observing whether those who make and review estimates are performing functions and using inputs as we understand they do  Inspecting documents supporting the controls used in the significant estimates  Observing whether the process owners, or others, act upon deviations from expectations As we confirm our understanding of the design of the relevant controls in the estimation SCOT, we consider how management determined whether:   Assumptions are appropriate (i.e., based on reasonable interpretations of present circumstances and the best available information)  Assumptions are consistent with those used in prior estimates or whether changes were needed because of changes in the environment or circumstances  Assumptions used are internally consistent with other assumptions used in the accounting estimate, other accounting estimates and with other aspects of the financial statements  Alternative assumptions may have been more appropriate  The external data used in making an estimate is relevant and reliable  The internal data used is relevant, complete and accurate  The IT applications are producing complete and accurate data through our understanding of the IT processes that support the IT applications. The method/model used to make the estimate is appropriate, and if it has been applied correctly Inquiry alone is not sufficient to confirm our understanding of, and evaluate, the design of the relevant controls. We review supporting documentation to confirm the results of the inquiries.  It may be appropriate to confirm our understanding of, and evaluate, the design of the relevant controls in conjunction with our testing of the operating effectiveness of the controls.  Substantive only strategy When we take a substantive only strategy for an estimation SCOT, we determine whether our understanding of the estimation SCOT is appropriate based on our substantive procedures. We may perform our procedures to identify and evaluate controls that address significant risks and the WCGWs over the preparation and posting (i.e., initiation, authorization and recording) of journal entries at the same time as our substantive procedures (refer to  ESTIMATES 10.1). However, as described in  ESTIMATES 7, we consider the following when determining whether our understanding of the estimation SCOT may need to occur earlier in the audit:</not_applicable_for_pcaob>\\n', '# STRATEGY: Audit strategy and CRA [effective for audits of periods ending before 15 December 2022]\\n  ##  4 Design an audit strategy to respond to risks of material misstatement at the assertion level\\n   ###  4.3 Substantive only strategy\\n  When we are not required to or we decide not to take a controls reliance strategy, we consider whether substantive procedures alone will provide sufficient appropriate audit evidence to address the identified risks of material misstatement for each relevant assertion of a significant account. If we do not believe sufficient information is available to design and execute substantive procedures that will reduce detection risk to an acceptably low level without mitigation from controls, we take a controls reliance strategy and follow the guidance in  STRATEGY 4.2. <REQUIREMENT> When we determine not to place reliance on controls relevant to the audit for a SCOT, we use the substantive only strategy.  ISA 315.20, AICPA AU-C 315.21</REQUIREMENT> When we use the substantive only strategy, we set control risk to “not rely on controls”. With a substantive only strategy, we obtain a sufficient understanding of the SCOT and supporting IT applications, from initiation to reporting, to determine the WCGWs to help design effective substantive procedures. Although we may believe a substantive only strategy is appropriate, we consider our approach for testing IPE and any risks introduced to the SCOT by IT processes. When we are relying on financial or non-financial data points that we cannot validate from substantive procedures alone, we follow the guidance in  IPE 4.5.3.  When we assess control risk as “not rely on controls” for an assertion, we obtain audit evidence  only from substantive procedures.   When we reach the conclusion to assess control risk as “not rely on controls” for an assertion, we combine our control risk assessment with our inherent risk assessment to arrive at a combined risk assessment of either moderate or high, as further described in  STRATEGY 5. We then consider the nature, timing, and extent of substantive evidence necessary to persuade us that there is no material misstatement. As we design our substantive procedures, we acknowledge that the nature and extent of substantive evidence is increased as detection risk becomes our only means to address the inherent risk at the assertion level.  ###  4.4 SCOTs with special circumstances\\n  The design of our audit strategy considers the relationship between control risk and detection risk; however, we also consider if the related SCOT includes any special circumstances that need to be addressed. We take additional action as part of understanding the SCOTs for:   SCOTs for which substantive procedures alone may not be sufficient ( SCOTS 5.1)  Related party relationships and transactions ( SCOTS 5.2)   Estimation SCOTs ( ESTIMATES 7)  Significant risks ( SCOTS 5.4)  Journal entries, including those that post transactions from the sub-ledger to the general ledger ( SCOTS 5.5)\\n', '# STRATEGY: Risk assessment procedures, audit strategy and CRA [effective for audits of periods ending on or after 15 December 2022]\\n  ##  6 Design an audit strategy to respond to risks of material misstatement at the assertion level\\n   ###  6.4 SCOTs with special circumstances\\n  The design of our audit strategy considers the relationship between control risk and detection risk; however, we also consider if the related SCOT includes any special circumstances that need to be addressed. We take additional action as part of understanding the SCOTs for:    SCOTs for which substantive procedures alone may not be sufficient ( SCOTS 5.1)   Related party relationships and transactions ( SCOTS 5.2)    Estimation SCOTs ( ESTIMATES 7)   Significant risks ( SCOTS 5.3)   Journal entries, including non-standard journal entries used to record non-recurring, unusual transactions or adjustments ( SCOTS 5.4)\\n\\n', ' # IT and automated techniques FAQs\\n  ## AT-06 [10-21]: What is the effect of testing ITGCs and finding them ineffective when using automated techniques?\\n  As stated in  FAQ AT-05, the hxPSPs support a substantive only strategy for the significant accounts and related SCOTs with a moderate CRA, and we do not rely on controls over the SCOT or the IT processes.  When we determine that ITGCs are ineffective and the IT processes do not support the continuous operation of IT application controls or the production of complete and accurate IPE, we consider the effect on our audit strategy when using automated techniques.  IT 9 details alternative procedures that may be performed when ineffective ITGCs are identified. We add the appropriate risks related to the use of IT (i.e., WCGWs) to our understanding of the SCOT and design our audit procedures to address the risks accordingly. The risks related to a manage change process are likely to be covered by the use of the hxPSPs for the significant accounts addressed by the HxPSPs because the hxPSPs are designed to detect risks of material misstatement which includes when inappropriate modifications to IT applications have a material effect on the financial statements.  With respect to a manage access process, we understand the types of misstatements that could occur and whether our hxPSPs appropriately respond to these risks. For example, we could identify all entries recorded by the generic users and determine whether transactions recorded by such users are appropriate.  However, we recognize that a manage access process that does not trace actions performed by users who have extended rights may present higher risks that may not be detected by our hxPSPs. In such situations, we design appropriate procedures to address any remaining risk, which may include performing hxPSPs to a higher level of sensitivity, performing additional procedures using automated techniques or performing more traditional audit procedures.  We also consider the effect of the ineffective ITGCs or ineffective IT processes on that portion of our audit procedures that are not executed using automated techniques, including on those significant accounts, related SCOTs and IPE, where we do not apply the hxPSP audit programs (refer to  IT 9).', ' # Risk assessment using automated techniques - Tool-specific guidance [AMER]\\n  ## A.4. Understand the SCOTs\\n     Guidance: Obtain an understanding of SCOTs using automated techniques by understanding:   Sources of transactions and related double-entry accounting for routine and non-routine SCOTs (Procedure  A.4a)  Who is involved in the process – for routine SCOTs (Procedure  A.4b)  Profile of credit and debit activity over time – for routine SCOTs impacting the income statement (Procedure  A.4c) To understand the SCOTs, we use a combination of automated techniques, inquiry, observation and inspection to obtain information to identify WCGWs. As we obtain our understanding of SCOTs, we identify the points where information is, or ought to be, captured, transferred, or modified including how IT affects the SCOT, and consider what can go wrong. We do not attempt to identify everything that could go wrong within the process, but focus where:   There is a likelihood of occurrence of misstatement  The potential misstatement can result in a material misstatement to the significant account, either individually or when aggregated with other potential misstatements  We perform the procedures in steps  A.4a-A.4c, as described below, for the following SCOTs, at a minimum:   Affected by a significant risk, including fraud risk  Related to hxPSPs being performed  Accounts where the disaggregation factor is being used (Digital GAM  SUBSTANTIVE SAP 4)  Related to specific risks of management override ( FRAUD RESPOND 6)  For non-PCAOB audits: We have identified controls as lower risk controls ( Digital GAM CONTROLS 4.1) We perform these procedures for other SCOTs when they help understand unusual items or changes to processes identified during other procedures All of the procedures outlined in steps  A.4a-A.4c below are performed for each significant account impacted by a routine SCOT in scope based on the above guidance. Our automated techniques typically give us greater insight into routine SCOTs than they do into non-routine or estimation SCOTs due to the level of automation within the entity’s accounting process and the way transactions are recorded, processed and reported. Therefore, our procedures and documentation to understand the SCOTs depend on the nature of the significant account and related SCOTs:  Significant accounts impacted by routine SCOTs: Perform  A.4a through A.4c.  Significant accounts not impacted by routine SCOTs: Perform procedure  A.4a; further procedures are recommended but not required.   Section B: Perform the procedures listed. We do not need to retain the analyses used in our workpapers when our understanding is documented narratively – refer to  DOCUMENTATION. ', ' # ICFR FAQs [PCAOB-IA only]\\n  ## Understanding the significant classes of transactions (SCOTs) and performing walkthroughs\\n  ### IC-01 [03-16]: Is performing a walkthrough sufficient to document our understanding of the SCOT?\\n  No. Thoroughly understanding the flow of transactions for the SCOTs to identify the risks of material misstatement for the accounts and assertions affected by these SCOTs (i.e., what can go wrongs (WCGWs)) is required by our professional standards and is fundamental to an effective audit of the financial statements and ICFR. We generally document our understanding of the SCOT separately from the walkthrough template in a narrative or a flow chart (or in combination). Our walkthrough is then used to confirm our understanding of the SCOT by selecting a transaction to walk through the various critical paths in the process (e.g., if different types of revenue transactions are processed differently, our understanding of the revenue SCOT would identify these different processes and our walkthroughs would include a walkthrough of each of the different types of revenue transactions in order to confirm our understanding). For estimation processes, the walkthrough may be documented differently as there may not be an individual transaction to walk through; however, we still perform procedures to confirm our understanding of the flow of the estimation process from initiation to reporting in the general ledger. Refer to  SCOTS 2,  SCOTS 7 and  CONTROLS 3 for additional guidance.  ### IC-02 [03-16]: Why does EY GAM require us to obtain an understanding of the SCOT and identify the WCGWs?\\n  Understanding the flow of transactions and risks of material misstatement is a fundamental requirement of PCAOB Auditing Standard 2201,  An Audit of Internal Control Over Financial Reporting That Is Integrated with An Audit of Financial Statements, (AS 2201) (f/k/a AS No. 5) and is critical to identifying all relevant risks of material misstatement in the flow of transactions and those controls that, individually or in combination, mitigate the risks within the SCOT. If we fail to focus on all important aspects of the flow of transactions or the WCGWs are generic or not aligned with the risks in the process, we may fail to identify whether a relevant control exists and thus our evaluation of the effectiveness of the design of the entity’s controls may not be appropriate (i.e., we may fail to determine that identified controls do not mitigate the risks). Further, if we do not understand how the SCOT operates and the WCGWs, we may not identify and test the appropriate controls or we may not design effective tests of those controls. Refer to  SCOTS 2,  SCOTS 3, and  SCOTS 4 for additional guidance.', ' # EY Private FAQs [US]\\n  ## SCOTS\\n   ### How is the Probing Inquiries approach different than the top sheet approach for a non-PIE audit? [29 July 2021]\\n  The  top sheet approach allows us to carryforward our prior period documentation for a recurring audit after we have obtained a current period transaction to confirm whether there have been any changes in the critical path from the prior period. The Probing Inquiries approach uses inquiries to confirm our understanding of the critical path for each routine and non-routine SCOT every period and to identify whether isolated or significant changes have occurred from the prior period based on the results of our inquiries.  For non-PIEs, when we identify no changes in the critical path of the SCOT from the prior period, we document the results of our probing inquiries in the Form 280-Post  SCOTs form and carry forward our prior-period documentation of the critical path (i.e., the tracing of a transaction), including the related prior-period walkthrough procedures evidence to satisfy the requirement of confirming our understanding of the SCOT. When using probing inquiries and have identified no change, we are not required to select and obtain evidence for a current period transaction unless we have already carried forward our transaction for two audit periods. The carry forward period of the ticked and tied transaction is the same under both approaches.  When, as a result of our probing inquiries, we identify changes in the critical path of the SCOT, we revise our process documentation, as necessary, including what can go wrongs (WCGWs) and, when applicable, the identified relevant controls. We evaluate the nature and extent of changes (i.e., isolated or pervasive) and perform procedures to update our understanding. When pervasive changes in a SCOT are identified, our procedures include tracing a transaction through the critical path. Refer to  SCOTS 7.2 for additional guidance. The use of probing inquiries does not change the requirements for procedures we perform to evaluate the design and determine the implementation of controls, including the requirements in  SCOTS 5 for special circumstances. We do not use probing inquiries to replace our tracing of a transaction through the critical path when we are performing an initial audit, or when we identify a new or non-recurring SCOT on a recurring audit. Refer to “What are the requirements for completing the Probing Inquiries approach?“ above for further details.  ### Are there any guidelines to how many periods in a row we can carry forward evidence for a non-PIE audit? [29 July 2021]\\n  At a minimum, we update our documentation that supports tracing a transaction along its critical path every three audit periods (i.e., we may carry forward prior period evidence for two audit periods after obtaining the original transaction documentation in period 1) which is consistent with the top sheet approach.  In the year of adoption, if we have already carried forward a walkthrough for two periods under the top sheet approach, we will need to select a current period transaction and complete a walkthrough of the critical path using Form 280-Post  SCOTs form (Canvas form).  This form retains content from prior period upon EY Canvas rollforward allowing for easy updating.']",
    'what are the additional procedures we do we a particular SCOT includes significant estimates',
    'What should I do if my population does not contain similar items',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Training Details

Training Dataset

json

• Dataset: json
• Size: 27 training samples
• Columns: positive and anchor
• Approximate statistics based on the first 27 samples:

  	positive	anchor
  type	string	string
  details	min: 512 tokens mean: 512.0 tokens max: 512 tokens	min: 7 tokens mean: 17.11 tokens max: 38 tokens

• Samples:

  positive	anchor
  ['# SUBSTANTIVE TOD: Tests of details\n ## 2 Determining the extent of our tests of details\n ### 2.2 Key items\n #### 2.2.1 Determining scopes and thresholds for key items\nIn some cases a population contains transactions of high value but do not represent higher risk or are not considered unusual. In such cases, we may not identify these transactions as key items but would still need to address the risk of material misstatement in the population through other procedures.', '# SUBSTANTIVE TOD: Tests of details\n ## 2 Determining the extent of our tests of details\n ### 2.2 Key items\n #### 2.2.1 Determining scopes and thresholds for key items\n Key items selected may include the following: Items with certain qualitative characteristics. We select and examine items that exhibit specific characteristics because they are more likely to contain material misstatements. These are items that are suspicious, unusual, have a higher risk of material misstatement, or have a history of misstatement. Refer to SUBSTANTIVE EXECUTE 1.2.\nHigh value items. We may select and examine specific items over a certain monetary amount within a population because they are of high value and therefore individually significant because of their size as compared to other items in the population. When designing our substantive procedures, we exercise judgment in setting the scopes and thresholds used to select key items. This is different from judgmentally determining sample sizes.', ' # SAMPLE: 4 Sampling for tests of details\n ## 4.8 Evaluate and conclude on the sample results\n ### 4.8.2 Misstatements in key items\n If we identify a misstatement in our key items, we treat it as a factual misstatement. We do not project misstatements identified in key items as key items are not representative of the population. ISA 450.6, ISA 530.14, AICPA AU-C 450.06, AICPA AU-C 530.13, PCAOB AS 2315.26, PCAOB AS 2810.14 If we find misstatements in key items, we consider: The reason for the misstatement in the key item and determine whether this misstatement could occur within the whole population, including other key items Whether the design of our testing performed on the remaining population (i.e., other than key items) would identify whether errors similar in nature and type to the ones we discovered in our key item testing were also present in the remaining population When our testing on the remaining population is designed to identify similar errors, and we identify such errors when performing our testing, we determine if the errors are the result of a representative error or common feature in accordance with SAMPLE 4.8.1a When our testing is not designed in a manner that would identify similar errors, we consider the need to obtain additional audit evidence in accordance with SAMPLE 4.8.1b to address the risk in the remaining population The effect of a misstatement in key items on our CRA, particularly on our control risk assessment (as misstatements may be an indication of ineffective controls) Misstatements in key items are posted to the summary of audit differences as factual misstatements', '# SUBSTANTIVE TOD: Tests of details\n ## 2 Determining the extent of our tests of details\n ### 2.2 Key items\n #### 2.2.3 Identifying high value key items for liability and expense accounts\n We generally do not identify key items based on a quantitative testing threshold for liability or expense accounts as our primary concern is with their completeness, and not their existence. Therefore, our criteria for selecting items to test may depend on other factors (e.g., we may select a sample of suppliers to send creditors confirmation requests based on the value of the purchase activity over the entire audit period). If we identify key items based on a testing threshold for liability or expense accounts to test for classification or valuation, we may use the same testing thresholds as for asset accounts.\nHowever, if we design a test for completeness that uses a reciprocal population (e.g., cash disbursements after the interim or balance sheet date), we reduce our testing thresholds to allow for the aggregation of small misstatements. Our testing for unrecorded liabilities may include the examination of expense postings or checks issued subsequent to the balance sheet date over a testing threshold to determine whether the related liabilities are recorded in the appropriate period (i.e., the population for detecting such understatements involves transactions that meet the criteria for being omitted from the account being tested).', '# SUBSTANTIVE TOD: Tests of details\n ## 2 Determining the extent of our tests of details\n ### 2.2 Key items\n #### 2.2.1 Determining scopes and thresholds for key items\n Key items are those that have a higher likelihood of misstatement, such as: Unusual items or transactions - refer to SUBSTANTIVE EXECUTE 1.2 Transactions recorded at or near the balance sheet or interim reporting date Related party transactions Higher value or older items as compared to other items in the population We use professional judgment and decide to select key items from a population based on factors such as: PCAOB AS 1105.25, PCAOB AS 2315.21 Our understanding of the entity’s business Our assessment of inherent risk and control risk Characteristics of the population being tested The identification of key items starts with the examination of the population to determine which transactions with specific characteristics have a higher risk of material misstatement and have different characteristics from the rest of the population. Because different risks may exist for the income statement and the balance sheet, our considerations of qualitative factors and thresholds for key items may vary across significant accounts. We use our understanding of the business and the results of analytical procedures to help identify the entity’s key customers and those whose transactions with the entity represent greater risks (e.g., a key customer whose credit rating has recently been downgraded).\n', '# SAMPLE: 4 Sampling for tests of details\n ## 4.5 Determine sample size\n ### 4.5.2 Judgmental sample size determination\n When using judgmental sampling, we apply professional judgment to determine sample sizes for tests of details. We consider factors such as our understanding of the entity’s business, preliminary assessments of inherent and control risks and the characteristics of the population being tested. As the purpose of sampling is to provide a reasonable basis for us to draw conclusions about the population from which the sample is selected, it is important that we judgmentally determine an appropriate sample size and then select sampling units without bias so that they are representative of the population. Conclusions can then be drawn about the population on a judgmental basis. There are many judgments that enter into the determination of sample sizes; accordingly, when using judgmental sampling there is no exact answer as to what the correct sample size is for a particular test. We consider the factors in SAMPLE 4.4 for tests of details and apply professional judgment in determining our sample size. We also consider the amount of assurance required from our sample in view of our CRA and the audit evidence obtained from our substantive procedures, including primary substantive procedures. The Audit Risk Model discussed in STRATEGY 4 [ Pre Dec 2022: STRATEGY 1.2] provides guidance on the relationship between inherent risk, control risk and detection risk. When determining a sample size judgmentally, we document the basis on which we determined the judgmental sample to be appropriate. In addition, we document the factors that influence our sample size ( SAMPLE 4.4) which are: The CRA TE compared to the population value that we are testing The level of audit evidence we have obtained or expect to obtain from other substantive procedures Our expectation of misstatements', ' # SUBSTANTIVE TOD: Tests of details\n ## 2 Determining the extent of our tests of details\n ### 2.2 Key items\n #### 2.2.1 Determining scopes and thresholds for key items\n If a very large number of key items are identified, that is generally an indicator that the quantitative threshold is too low and those items may not be different enough from other items in the population to be considered key items. We analyze the characteristics of the population in order to identify the items that are individually significant based on size, contain a relevant event and condition or are otherwise unusual and, therefore, are not representative of the remaining population. We keep in mind that quantitative thresholds are only one of the factors we consider to identify key items. If we set our testing threshold above TE, we consider documenting our rationale for doing so.', ' # SAMPLE: 4 Sampling for tests of details\n ## 4.8 Evaluate and conclude on the sample results\n ### 4.8.4 Conclude on sample results\n #### 4.8.4b Conclude on random samples\n As the total uncorrected misstatement approaches TE, the likelihood that the population is materially misstated increases. We consider the following general guidelines: If the total uncorrected misstatement is low relative to tolerable error (e.g., less than 30% of TE) and our other procedures support our conclusions, we typically would not perform any additional procedures and can conclude that the sample results support the conclusion that the population is not misstated by more than TE. If the total uncorrected misstatement is large relative to TE (e.g., exceeds 50% of TE), the sample results most likely do not support the recorded amount of the population. It may not be appropriate to continue with our planned audit strategy as the total uncorrected misstatement is large in comparison to TE. Therefore, in these situations we request management to investigate identified misstatements and the potential for further misstatements in the account balance. If the total uncorrected misstatement is between 30% to 50% of TE, it is likely that further work needs to be performed to determine if we can conclude that the sample results support the conclusion that the population is not misstated by more than TE. We exercise professional judgment to determine if the sample results support our conclusion. In making this determination, we consider: Sampling risk (i.e., the risk that such a result may be obtained even though the true monetary misstatement for the population exceeds TE) Other misstatements identified during the audit The results of our key item testing The persuasiveness of the sample results considering our expectations (e.g., errors in processing were expected) Other procedures we have performed that provide audit evidence about the population In this case, we consider whether additional procedures are required to allow us to determine that the sample provides a reasonable basis for conclusions about the population that we tested. In making this determination, we consider if we have obtained coverage from key item testing. Although our tests of key items are not representative of the population, we consider the results of our tests of key items in determining if the sample provides a reasonable basis for conclusions about the population (i.e., if we find errors in key items, we may be more likely to find errors in the remainder of the population). In situations where we have not performed any key item testing (i.e., no key item coverage), we will be less likely to be able to reach a conclusion that the sample provides a reasonable basis for conclusions about the population that we tested. Additional procedures we may perform include: Requesting that management investigate identified misstatements and the potential for further misstatements in the account balance Expanding our key item coverage by reducing our testing threshold Expanding our representative sample to obtain additional audit evidence (expanding our sample is only a viable option if our hypothesis remains that the account balance is substantially recorded correctly after understanding the cause of any misstatements) Switching to a different sampling strategy (i.e., variables estimation sampling) In all situations, we request management to correct any key item misstatements (factual misstatements) and we post the projected misstatement above our SAD nominal amount to our summary of audit differences. We follow the guidance in MISSTATE 4.11 to evaluate and conclude on uncorrected misstatements.', ' # APM EDUCATION 2.3 Policy FAQs\n ## 7. Americas - I completed foundation learning in EY Leads in 2018/2019. Why do my completions not show in SuccessFactors?\n While EY Leads learning history was migrated into SuccessFactors there were variations in how the item details were carried across. As a result, there may be instances where EY Leads completions are not credited in the SuccessFactors CBJ. If you completed a foundation item in EY Leads in 2019, but you are not seeing a credit for it in SuccessFactors, then you should contact your local learning team to request an exemption from the relevant foundation learning item. If you had previously complied with US for non-US jurisdiction requirements but did not comply with 2019 cycle requirements, then the gap year guidance is applicable. (September 2020)', ' # SQM ROLES: Assignment of system of quality management roles and responsibilities\n ## Appendix 1 - Illustrative Country SQM representation language\n Country Independence Leader I confirm, to the best of my knowledge and belief, that: I understand the objectives and requirements of International Standard on Quality Management 1 and have fulfilled my SQM-related responsibilities. I have reviewed and agree with any findings or deficiencies in design, implementation and operation of the Country Independence-owned SQM key controls. I have reviewed and approved the effectiveness of the actions included in any Country Independence SQM Quality Improvement Plans and reviewed progress against the actions included in the Country Independence SQM Quality Improvement Plans. [Signature] Individual(s) with operational responsibility I confirm, to the best of my knowledge and belief, that: I understand the objectives and requirements of International Standard on Quality Management 1 and have fulfilled my SQM-related responsibilities. I understand all attestations made by individuals that are leaders of an: (1) EY Function; (2) Service Line functions that enable or monitor quality; and (3) EY Service Lines and made inquiries or performed other procedures, as appropriate, related to any comments made in their attestations. For control(s) for which I am the Control Owner, I understand and have fulfilled my Control Owner responsibilities and completed my Control Owner certification. I have reviewed and agree with the results of the SQM monitoring activities, including the analysis of findings to identify deficiencies (if applicable). I have reviewed and approved the effectiveness of the actions included in any SQM Quality Improvement Plans and reviewed progress against the actions included in the SQM Quality Improvement Plans. I have reviewed the Country SQM Annual Evaluation Report in its entirety. I recommend the SQM Annual Evaluation Conclusions for the Member Firms as stated in the table above. [Signature] Individual(s) with operational responsibility for monitoring I confirm, to the best of my knowledge and belief, that: I understand the objectives and requirements of International Standard on Quality Management 1 and have fulfilled my SQM-related responsibilities. I understand all attestations made by individuals that are leaders of a: (1) EY Function; (2) Service Line functions that enable or monitor quality; and (3) EY Service Lines and made inquiries or performed other procedures, as appropriate, related to any comments made in their attestations. For control(s) for which I am the Control Owner, I understand and have fulfilled my Control Owner responsibilities and completed my Control Owner certification. I have reviewed and approved the results of the SQM monitoring activities, including the analysis of findings to identify deficiencies (if applicable). I have reviewed and monitored the effectiveness of the actions included in any SQM Quality Improvement Plans. I have reviewed the Country SQM Annual Evaluation Report in its entirety. I concur with the SQM Annual Evaluation Conclusions for the Member Firms as stated in the table above. [Signature]']	How do I identify a key item in my population?
  ['# SAMPLE: 4 Sampling for tests of details\n ## 4.8 Evaluate and conclude on the sample results\n ### 4.8.1 Determine the nature and cause of misstatements\n #### 4.8.1a Misstatements that are not representative of the population\n If we conclude, based on our investigation of the nature and cause of a misstatement, that the misstatement is not representative of the population, we design an alternative audit strategy to test the items that are not representative of the population. Our audit strategy will depend on whether the misstatement contains a common feature or is an anomaly. Misstatements that contain a common feature If we determine the identified misstatements contain a common feature (e.g., type of transaction, location, product line or period of time), we identify all items in the population that possess the common feature and extend audit procedures to those items.\nWhen it is possible to contain a misstatement with a common feature to a distinct sub-population, we divide the initial sampling population into two sub-populations: one containing the items with the common feature; the second containing the remainder of the population. The two sub-populations are then evaluated separately. For the sub-population containing the remainder of the population, we determine whether we have sufficient appropriate audit evidence to conclude. If we do not identify any further misstatements, we may be able to conclude the remaining sample is sufficient and representative of the sub-population. If we identify other misstatements in either population, we determine if the misstatements are representative of the initial sampling population, a common feature or an anomaly (refer to SAMPLE 4.8.1). If we determine the other misstatements are representative, we evaluate them across the initial sampling population. For the sub-population containing the common feature, we ask the entity to identify all items related to the common feature and take appropriate actions. If the entity identifies and corrects all the items related to the common feature, we perform procedures to verify that there are no further misstatements in the corrected sub-population. We decide on one of the following approaches to obtain sufficient appropriate audit evidence over the corrected sub-population: Test 100% of the items that contain the common feature Test a sample or key items from the sub-population Perform other substantive procedures (e.g., substantive analytical procedure) There may be situations when the entity decides on an alternative approach (i.e., instead of identifying and correcting all items that relate to the common feature) by performing an analysis of the cause of the misstatement. Based on the results of the analysis, the entity may: Conclude that the effect of the common feature is immaterial. In this case, we perform procedures to evaluate the appropriateness of the conclusions reached. Determine that it is sufficient to identify and correct only some of the items related to the common feature. In this case, we perform procedures to evaluate management’s approach, test the items corrected by the entity and evaluate the appropriateness of the conclusions reached. The approach we take depends on the facts and circumstances of the items that contain the common feature. However, we need to carefully evaluate the cause of the misstatement and how the entity has corrected it before concluding which approach for these items is the appropriate audit strategy. We keep in mind that such misstatements may be intentional and may indicate the possibility of fraud.', ' # SAMPLE: Appendix 1 Audit risk tables\n ## Appendix 1.1 Appropriateness of the ARTs\n The first decision in using the ARTs is whether they are appropriate to the type of audit procedures being performed. The ARTs are appropriate when we have a homogenous population (i.e., that the characteristics of each item in the population are similar and therefore the risk associated with the population is the same - refer to SAMPLE 4.3) and we want to draw a conclusion on the monetary value of that population. We use the ARTs to determine our sample size when we: Perform tests of details for existence Test the valuation of accounts We expect a low misstatement situation The ARTs are typically used for accounts that contain overstatement. However, in certain circumstances they can be used when our audit objective is to detect understatement. In order to use the ARTs for those accounts where our audit objective is to detect understatement, we use an independent related population (e.g., a complete population of cash disbursements subsequent to the accounting period end date or a complete population of supplier statements), to which we apply the ARTs to select our sample for testing, evaluate the overstatement of that population and thereby conclude on the understatement of the related population. We do not use the ARTs when: We are not testing a representative sample because our tests of key items, together with other substantive procedures, provide sufficient appropriate audit evidence for the assertion A threshold for selecting items is for a second or lower level of testing related to a given account Performing procedures at the entity’s request and a threshold is established for selecting items in areas that are not material to the financial statements We have assessed combined risk as high and we expect more than a few misstatements We are performing procedures on a sample of a sample ', ' # SAMPLE: 4 Sampling for tests of details\n ## 4.8 Evaluate and conclude on the sample results\n ### 4.8.2 Misstatements in key items\n If we identify a misstatement in our key items, we treat it as a factual misstatement. We do not project misstatements identified in key items as key items are not representative of the population. ISA 450.6, ISA 530.14, AICPA AU-C 450.06, AICPA AU-C 530.13, PCAOB AS 2315.26, PCAOB AS 2810.14 If we find misstatements in key items, we consider: The reason for the misstatement in the key item and determine whether this misstatement could occur within the whole population, including other key items Whether the design of our testing performed on the remaining population (i.e., other than key items) would identify whether errors similar in nature and type to the ones we discovered in our key item testing were also present in the remaining population When our testing on the remaining population is designed to identify similar errors, and we identify such errors when performing our testing, we determine if the errors are the result of a representative error or common feature in accordance with SAMPLE 4.8.1a When our testing is not designed in a manner that would identify similar errors, we consider the need to obtain additional audit evidence in accordance with SAMPLE 4.8.1b to address the risk in the remaining population The effect of a misstatement in key items on our CRA, particularly on our control risk assessment (as misstatements may be an indication of ineffective controls) Misstatements in key items are posted to the summary of audit differences as factual misstatements', ' # SUBSTANTIVE TOD: Tests of details\n ## 2 Determining the extent of our tests of details\n ### 2.2 Key items\n We select our key items by identifying transactions or balances that: ISA 500.10, PCAOB AS 2315.21 Meet qualitative criteria that indicate that they are more likely to contain material misstatements (either individually or in the aggregate), or Are individually important because of their size Selective examination of key items from an account balance or significant class of transactions (SCOT) is often an efficient means of obtaining audit evidence. Key items are separated from other items in the population for testing because of their qualitative characteristics or high value, as they are not, by definition, representative of the entire population. Key items do not constitute audit sampling because the results of procedures applied to key items cannot be projected to the entire population. We aggregate the evidence obtained from examining key items with additional substantive evidence addressing the same assertion, as described in SUBSTANTIVE EXECUTE 2.2, in order to conclude on the assertion. When our combined risk assessment (CRA) is ‘minimal’, examining key items may provide sufficient audit evidence so that there is little risk of the remaining items in the population containing a material misstatement. When our CRA is other than ‘minimal’, we may need additional audit evidence from other procedures to address the higher risk of material misstatement. We consider the need to obtain audit evidence regarding the remaining population (i.e., the full population less key items). The risk that the remaining population may contain a material misstatement is higher when the value of the remaining population increases relative to TE. When additional evidence is needed, we perform further substantive procedures (e.g., substantive analytical procedures or representative sampling) over the remaining population to allow us to reach a conclusion.', ' # 18.\tGeneral technology issues\n ## 18.7.\tA piece of evidence has disappeared from my engagement. What should I do?\n Please contact the Help Desk, as there could be many reasons for this problem. Do not attempt to solve the issue on your own as this could cause additional complications. (Nov 2015)', ' # 18.\tGeneral technology issues\n ## 18.6.\tI cannot see my changes in the current version of the document, what should I do?\n Please contact the Help Desk, as there could be many reasons for this problem. Do not attempt to make any further changes to the document or attempt to solve the issue on your own as this could cause additional complications. (Nov 2015)', ' # 11.\tConclude and archive\n ## 11.13.\t What should I do if I need to restore my engagement for AQR and I already have two versions restored?\n You will first need to delete one of the existing restored engagement. After that, you will be able to create a new restore. If you cannot delete one of the existing restored engagement, you should reach out to your Area Activation Network and they can assist you with getting a third restored engagement. (Nov 2019)', ' # 18.\tGeneral technology issues\n ## 18.8.\tI’m getting an error message on my screen that starts with “Error XXXX” (where XXXX represents a number). What should I do?\n You should contact the Help Desk if you get any error messages in EY Canvas. Prior to contacting the Help Desk, you should attempt to recreate your steps of how the error came about, and when possible take screenshots of the error. This will help the technician determine the cause of the error. (Nov 2015)', " # ISA 530 - AUDIT SAMPLING\n ## Definitions\n 5. For purposes of the ISAs, the following terms have the meanings attributed below: (a) Audit sampling (sampling) – The application of audit procedures to less than 100% of items within a population of audit relevance such that all sampling units have a chance of selection in order to provide the auditor with a reasonable basis on which to draw conclusions about the entire population. (b) Population – The entire set of data from which a sample is selected and about which the auditor wishes to draw conclusions. (c) Sampling risk – The risk that the auditor's conclusion based on a sample may be different from the conclusion if the entire population were subjected to the same audit procedure. Sampling risk can lead to two types of erroneous conclusions: (i) In the case of a test of controls, that controls are more effective than they actually are, or in the case of a test of details, that a material misstatement does not exist when in fact it does. The auditor is primarily concerned with this type of erroneous conclusion because it affects audit effectiveness and is more likely to lead to an inappropriate audit opinion. (ii) In the case of a test of controls, that controls are less effective than they actually are, or in the case of a test of details, that a material misstatement exists when in fact it does not. This type of erroneous conclusion affects audit efficiency as it would usually lead to additional work to establish that initial conclusions were incorrect. (d) Non-sampling risk – The risk that the auditor reaches an erroneous conclusion for any reason not related to sampling risk. (Ref: Para A1) \xa0 (e) Anomaly – A misstatement or deviation that is demonstrably not representative of misstatements or deviations in a population. (f) Sampling unit – The individual items constituting a population. (Ref: Para A2) (g) Statistical sampling – An approach to sampling that has the following characteristics: (i) Random selection of the sample items; and (ii) The use of probability theory to evaluate sample results, including measurement of sampling risk. A sampling approach that does not have characteristics (i) and (ii) is considered non-statistical sampling. (h) Stratification – The process of dividing a population into sub-populations, each of which is a group of sampling units which have similar characteristics (often monetary value). (i) Tolerable misstatement – A monetary amount set by the auditor in respect of which the auditor seeks to obtain an appropriate level of assurance that the monetary amount set by the auditor is not exceeded by the actual misstatement in the population. (Ref: Para A3) (j) Tolerable rate of deviation – A rate of deviation from prescribed internal control procedures set by the auditor in respect of which the auditor seeks to obtain an appropriate level of assurance that the rate of deviation set by the auditor is not exceeded by the actual rate of deviation in the population.", ' # 18.\tGeneral technology issues\n ## 18.4.\tWhat do I do if I don’t have the right version of Doc Helper on my computer?\n The most recent version of Document Helper can be obtained from the ServiceNow Service Management website ( https://eyt.service-now.com/ey/). (May 2020)']	What should I do if my population does not contain similar items
  ['# 8-11 Update and evaluate controls\n ## 10 Evaluate deficiencies in internal control\n We determine whether, on the basis of the audit work performed, we have identified one or more deficiencies in internal control. ISA 265.7, AICPA AU-C 265.08 If we conclude that a control is not operating effectively and a control deficiency exists, we evaluate it to determine if it constitutes a significant deficiency (or material weakness if required to be communicated in the jurisdiction) that we need to communicate to management and those charged with governance. Refer to SOCD\xa01. We may identify deficiencies in: Design, when we confirm our understanding of controls through performing walkthrough procedures Operation, as we execute our tests of controls A deficiency in design exists when either: A control necessary to address a WCGW or IT risk is missing An existing control is inappropriately designed so that, even if it operates as designed, its objective will not be met (e.g., it will not prevent a misstatement occurring on a timely basis) A deficiency in operation exists when either: An appropriately designed control does not operate as intended The person performing the control does not possess the necessary authority, capabilities, competence or objectivity to do it effectively Generally, control exceptions indicate a deficiency in internal control, except in the following limited situations: Upon further evaluation, we determine it was not a deviation from the intended functioning of the control (i.e., we misinterpreted the exception) We obtain sufficient appropriate audit evidence by performing additional procedures, beyond inquiry, to conclude that the rate of exception in the functioning of the control is acceptable based on our understanding of its design (e.g., we extend the sample size for a random control exception, refer to CONTROLS\xa08.2a) \nFor integrated audits, in the limited circumstances when we conclude that a control exception is not considered a deficiency, we generally perform additional testing to conclude that the control is operating effectively, or the rate of exception is acceptable. PCAOB AS 2201.48\n\nFor integrated audits, we consider accumulating control exceptions determined not to be control deficiencies to facilitate audit executive review and determine that we adequately evaluated and documented each control exception. Such accumulation may include the control description, related SCOT, business unit/location, description of the exception and explanation why it is not a control deficiency. We include exceptions that we identified from our independent tests of controls, or from testing performed by management, internal audit or others.\nThe existence of effective compensating controls does not remove a control deficiency.', ' # SOCD : Evaluate control deficiencies\n ## Purpose\n We evaluate control deficiencies identified through our audit procedures to determine if they constitute a significant deficiency. We evaluate both identified deficiencies in the design of controls and in the operation of controls. Design deficiencies include those that are identified related to the appropriateness of the entity’s policies in each of the components of the system of internal control (refer to ELC [ Pre Dec 2022: ELC] and FSCP 6). In some jurisdictions, for example when reporting as an integrated audit under PCAOB standards, we also determine whether the control deficiency is a material weakness. Throughout EY GAM, we discuss identifying significant deficiencies and use the phrase “or material weakness if required to be communicated in the jurisdiction” to address these situations. We refer to local requirements when determining whether a control deficiency is a material weakness. We communicate significant deficiencies to management and those charged with governance, and when applicable in the jurisdiction, we communicate separately material weaknesses. For guidance on determining whether we have identified a control exception and whether it is a deficiency refer to CONTROLS 8 and CONTROLS 10. This topic provides requirements and guidance to determine whether a control deficiency constitutes a significant deficiency (or material weakness if required to be communicated in the jurisdiction).', '# SOCD : Evaluate control deficiencies\n ## 1 Significant deficiencies (or material weaknesses, if required to be communicated in the jurisdiction)\n #### 1.1 Indicators of a material weakness\n In addition to the quantitative and qualitative factors discussed in SOCD 1, we consider indicators of material weaknesses in internal control when evaluating whether a control deficiency is a material weakness. Indicators of material weaknesses in internal control include: Identification of fraud, whether or not material, on the part of senior management Restatement of previously issued financial statements to reflect the correction of a material misstatement due to fraud or error Identification by us of a material misstatement of the financial statements under audit in circumstances that indicate that the misstatement would not have been detected by the entity’s internal control, and Ineffective oversight of the entity’s financial reporting and internal control by those charged with governance PCAOB AS 1305.05 When we evaluate the significance of a deficiency, or combination of deficiencies, in internal control over financial reporting, we determine the level of detail and degree of assurance that would satisfy prudent officials in the conduct of their own affairs that they have reasonable assurance that transactions are recorded as necessary to permit the preparation of financial statements in accordance with the applicable financial reporting framework. PCAOB AS 2201.70, AICPA AU-C 265.10 If we determine that the deficiency, or combination of deficiencies, would prevent prudent officials in the conduct of their own affairs from concluding that they have reasonable assurance that transactions are recorded as necessary to permit the preparation of financial statements in accordance with the applicable financial reporting framework, then we deem the deficiency, or combination of deficiencies, to be an indicator of a material weakness. PCAOB AS 2201.70, AICPA AU-C 265.10 \nGiven the high degree of judgment required, we are required to consult with the Regional PPD if any of the above circumstances are noted because they are indicators that a material weakness in internal control over financial reporting exists. Such consultation includes the effects on our report and the implications on client continuance. EY Policy These matters require consultation with the Regional PPD and should not be delegated to other partners within the PPD group assigned to the Region.\n\nThe following diagram can be used to assist in evaluating deficiencies as it graphically depicts the evaluation process required under PCAOB AS 2201:\n', '# SOCD : Evaluate control deficiencies\n ## 1 Significant deficiencies (or material weaknesses, if required to be communicated in the jurisdiction)\nFor integrated audits, we include consideration of the entity’s interim financial statements in our evaluation of a material weakness.\nA control deficiency alone may not represent a significant deficiency (or material weakness if required to be communicated in the jurisdiction). However, a combination of deficiencies affecting the same significant account or disclosure, or relevant assertion, may increase the risk of material misstatement and give rise to a significant deficiency (or material weakness). \nWe determine whether individual control deficiencies that affect the same significant account or disclosure, relevant assertion or component of internal control individually or in combination result in a material weakness. PCAOB AS 2201.65\n\nWe, management, the internal audit function or other third parties may identify control deficiencies. Regardless of who identifies the control deficiency, we consider all control deficiencies accumulated in our evaluation. Our evaluation requires significant professional judgment and involves audit executives. We do not accumulate deficiencies identified by others that do not affect the identified significant accounts and disclosures and relevant assertions.\nThe severity of a deficiency depends on: Whether there is a reasonable possibility that the entity’s controls will fail to prevent, or detect and correct, a misstatement in an account or disclosure The magnitude of the potential misstatement resulting from the deficiency or deficiencies The severity of a deficiency does not depend on whether a misstatement actually has occurred but on whether there is a reasonable possibility that the entity’s controls will fail to prevent or detect and correct a misstatement of an account balance or disclosure. Significant deficiencies and material weaknesses may exist even though we have not identified misstatements during the audit. The evaluation of whether a deficiency in internal control presents a reasonable possibility that a misstatement may occur may be made without quantifying the probability of occurrence as a specific percentage or range.', '# SOCD : Evaluate control deficiencies\n ## 1 Significant deficiencies (or material weaknesses, if required to be communicated in the jurisdiction)\n When we identify one or more control deficiencies, we determine, on the basis of the audit procedures performed, whether individually or in combination, they constitute a significant deficiency (or material weakness if required to be communicated in the jurisdiction) in internal control. ISA 265.8, AICPA AU-C 265.09 \nWe evaluate the severity of all identified control deficiencies to reach an overall conclusion as to whether control deficiencies, individually or in combination, are significant deficiencies or material weaknesses as of the date of management’s assessment. We are not required to search for deficiencies that, individually or in combination, are less severe than a material weakness. However, as part of our responsibilities for communication with those charged with governance, we consider whether there are any deficiencies, or combinations of deficiencies, that have been identified during the audit that are significant deficiencies (refer to SOCD 4). PCAOB AS 2201.62\n As defined in the ISAs, a significant deficiency in internal control is a deficiency, or a combination of deficiencies, in internal control that, in our professional judgment, is of sufficient importance to merit the attention of those charged with governance. As defined in US auditing standards, a significant deficiency in internal control is a deficiency, or a combination of deficiencies, in internal control that, in our professional judgment, is less severe than a material weakness, yet is of sufficient importance to merit the attention of those charged with governance. As defined in US auditing standards, a material weakness is a deficiency, or combination of deficiencies, such that there is a reasonable possibility that a material misstatement of the entity’s annual financial statements will not be prevented or detected on a timely basis.', " # 3E_3.5 Internal control report modifications [AICPA audits ending prior to 15 December 2021, All PCAOB audits]\n ## 3E_3.5 Non-Issuers Example 1 - Disclaimer of opinion on internal control over financial reporting- scope limitation\n Because of the matter described above, we were not able to obtain sufficient appropriate audit evidence to provide a basis for an audit opinion. However, a material weakness has been identified. A material weakness is a deficiency, or a combination of deficiencies, in internal control over financial reporting, such that there is a reasonable possibility that a material misstatement of the Company's financial statements will not be prevented, or detected and corrected, on a timely basis. If one or more material weaknesses exist, a company's internal control over financial reporting cannot be considered effective. The following material weakness has been included in the accompanying [ title of management's report]. [ Identify the material weakness described in management's report and include a description of the material weakness, including its nature and its actual and potential effect on the presentation of the entity's financial statements issued during the existence of the material weakness.] In addition, we include the following sentence in the Report on Financial Statements section: We considered the material weakness identified above in determining the nature, timing, and extent of audit procedures applied in our audit of the [ indicate year, e.g., “20X7”] financial statements, and this report does not affect such report on the financial statements. If, through the limited procedures we performed, we conclude that a material weakness exists that has not been identified by management, we indicate this fact in our report and we include a description of the material weakness: Material Weakness In performing our procedures, the following material weakness has been identified that was [ if management does not issue a report state, “not identified by management”, if management has issued a report but has not identified the material weakness state “not included in (title of management’s report)”]. [ Include a description of the material weakness, including its nature and its actual and potential effect on the presentation of the entity’s financial statements issued during the existence of the weakness.] A material weakness is a deficiency, or a combination of deficiencies, in internal control over financial reporting, such that there is a reasonable possibility that a material misstatement of the Company’s financial statements will not be prevented, or detected and corrected, on a timely basis. If one or more material weaknesses exist, a company's internal control over financial reporting cannot be considered effective. In addition, we include the following sentence in the Report on Financial Statements section: We considered the material weakness identified above in determining the nature, timing, and extent of audit procedures applied in our audit of the [ indicate year, e.g., “20X7”] financial statements, and this report does not affect such report on the financial statements. ", '# 8-11 Update and evaluate controls\n ## 8 Address control exceptions\n ### 8.2 Determine audit response to control exceptions\nWhen we have identified a control deficiency, before concluding on an appropriate audit response, we consider each of the following: The nature and severity of the deficiency (e.g., do the facts and circumstances indicate that the deficiency could be a significant deficiency (or material weakness if required to be communicated in the jurisdiction) if not remediated) The nature of the control (e.g., manual or automated control) and the nature and timing of the remediation (e.g., was the deficiency identified late in the year, or was it identified early in the year but management failed to remediate it earlier) Whether we believe the control is operating effectively as of period end and whether the evidence we would obtain from a lower extent of testing provides a basis for such a conclusion If a remediated control would otherwise represent a material weakness, the CAML is available to assist the team in developing the appropriate remediation testing strategy.\nIf the identified control deficiency has been remediated prior to period end, it will not be evaluated further as part of our period end assessment of the effectiveness of internal control over financial reporting. However, if the control deficiency has not been remediated and tested before period end, we further evaluate the control deficiency to determine if it constitutes a significant deficiency (or material weakness). Even if a control is successfully remediated before period end, we reconsider our financial statement audit strategy because the control was not operating effectively throughout the entire audit period.\n\n', ' # INTEGRATED AUDIT - Reports on internal control over financial reporting [effective for AICPA audits of periods ending on or after 15 December 2021]\n ## 5 Internal control report modifications\n ### 5.1 One or more material weaknesses exist\n #### 5.1.1 Effect of material weakness on prior periods\n Management may determine that a material weakness identified in the current year also was present in one or more previous assessment periods. When describing the potential effects on the financial statements issued during the existence of the material weakness, it is not necessary or appropriate to describe the effects of the material weakness on the earlier periods unless there also is a restatement of the prior year financial statements that coincides with the discovery of the material weakness. The annual assessment of internal control over financial reporting is not intended to be a comparative presentation and the current year assessment in effect supersedes the previous assessment. #### 5.1.2 Management unable to assess controls\n Management’s inability to assess certain controls over financial reporting that should have been included in management’s assessment represents a control deficiency in the control environment and monitoring components of internal control over financial reporting that we must evaluate as to significance. If the transactions or balances subject to controls that management is unable to assess are material to the entity’s financial statements, we ordinarily would determine that this control deficiency represents a material weakness. Refer to INTEGRATED AUDIT 5.3.1 for further requirements and guidance. #### 5.1.3 Management remediates material weakness too late to test\n We may encounter a situation where management identifies and remediates a material weakness too late for them or us to test the new control prior to the filing date. We would conclude that, absent evidence that the control has been remediated, the material weakness still exists. If management concludes that internal control is ineffective and appropriately identifies the material weakness in its report, we issue an adverse opinion on the effectiveness of internal control over financial reporting. Refer to Appendix 2.2.4 – INTEGRATED AUDIT – Illustration 9 for an example report.', ' # 3.6 Adverse opinion on the effectiveness of internal control over financial reporting [AICPA audits ending prior to 15 December 2021, All PCAOB audits]\n ## 3.6.1 General guidance\n ### 3.6.1.1 Effect of material weakness on prior periods\n Management may determine that a material weakness identified in the current year also was present in one or more previous assessment periods. When describing the potential effects on the financial statements issued during the existence of the material weakness, it is not necessary or appropriate to describe the effects of the material weakness on the earlier periods unless there also is a restatement of the prior year financial statements that coincides with the discovery of the material weakness. The annual assessment of internal control over financial reporting is not intended to be a comparative presentation and the current year assessment in effect supersedes the previous assessment. ### 3.6.1.2 Management unable to assess controls\n Management’s inability to assess certain controls over financial reporting that should have been included in management’s assessment represents a control deficiency in the control environment and monitoring components of internal control over financial reporting that we must evaluate as to significance. If the transactions or balances subject to controls that management is unable to assess are material to the entity’s financial statements, we ordinarily would determine that this control deficiency represents a material weakness. Refer to RM 3.5.3.2 Scope limitations imposed by circumstances for further requirements and guidance. ### 3.6.1.3 Management remediates material weakness too late to test\n We may encounter a situation where management identifies and remediates a material weakness too late for them or us to test the new control prior to the filing date. We would conclude that, absent evidence that the control has been remediated, the material weakness still exists. If management concludes that internal control is ineffective and appropriately identifies the material weakness in its report, we issue an adverse opinion on the effectiveness of internal control over financial reporting.\xa0Refer to [Issuers] RM 3E_3.6 Issuers Example 1 [Issuers] [Non-Issuers] RM 3E_3.6 Non-Issuers Example 1 [Non-Issuers] for an example report.', ' # Amendments to Rules Regarding Management's Report on Internal Control Over Financial Reporting, Release Nos. 33-8809; 34-55928; FR-76, 6/27/2007\n ## I. BACKGROUND\n 3. Final Rule After consideration of the comments received, we have determined that it is appropriate for the Commission's rules to include the definition of material weakness since it is an integral term associated with Sarbanes-Oxley and the Commission's\xa0 implementing rules. Management's disclosure requirements with respect to ICFR are predicated upon the existence of a material weakness; therefore, we agree with the commenters' suggestion that our rules should define this term, rather than refer to auditing literature. As a result, we are amending Exchange Act Rule 1 2b-2 and Rule 1- 02 of Regulation S-X to define the term material weakness. We have decided to adopt the material weakness definition substantially as proposed. The Commission has determined that the proposed material weakness definition appropriately describes those conditions in ICFR that, if they exist, should be disclosed to\xa0 investors and should preclude a conclusion that ICFR is effective. Therefore, our final rules define a material weakness as a deficiency, or a combination of deficiencies, in ICFR such that there is a reasonable possibility that a material misstatement of the registrant's annual or interim financial statements will\xa0 not be prevented or detected on a timely basis. Exchange Act Rule 1 2b-2 and Rule 1-02(p) of Regulation S-X.\xa0\xa0 We anticipate that the PCAOB's auditing standards will also include this definition of material weakness. After consideration of the proposed alternatives to the "reasonable possibility" standard in the proposed definition of material weakness, we decided not to change the proposed standard. Revisions that have the effect of increasing the likelihood\xa0 (that\xa0 is, risk) of a material misstatement in a company's financial reports that can exist before being disclosed could give rise to questions about the meaning of a disclosure that ICFR is effective and whether the threshold for "reasonable assurance" is\xa0 being\xa0 lowered. Moreover, we do not believe improvements in efficiency arising from revisions to the likelihood element would be significant to the overall ICFR evaluation effort, due, in part, to our view that the effort evaluating deficiencies would be similar\xa0 under the alternative standards (for example, "reasonable possibility" as compared to "reasonable likelihood"). Lastly, we do not believe the volume of material weakness disclosures, which has declined each year since the initial implementation of Section 404 of SarbanesOxley, is too high such that investors would benefit from a reduction in disclosures that would result from a higher likelihood threshold. Regarding the reference to interim financial statements in the definition of material weakness, while we believe annual materiality considerations are appropriate when making judgments about the nature and extent of evaluation procedures, we believe\xa0 that the judgments about whether a control is adequately designed or operating effectively should consider the requirement to provide investors reliable annual and quarterly financial reports. Moreover, if management's annual evaluation identifies a\xa0 deficiency that poses a reasonable possibility of a material misstatement in the company's quarterly reports, we believe management should disclose the deficiency to investors and not assess ICFR as effective. As such, we have not removed the reference to\xa0 interim financial statements from the definition of material weakness. In response to the comments regarding the need for the Commission to define the term "significant deficiency," we are seeking additional comment on a definition of that term as part of a separate release issued in the Federal Register.']	if we identify an exception, hoow can we determine that the exception falles under deficiency, weakness or material weakness

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          128,
          64
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• num_train_epochs: 1
• warmup_ratio: 0.1
• fp16: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• learning_rate: 5e-05
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 1
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.1
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• save_safetensors: True
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• no_cuda: False
• use_cpu: False
• use_mps_device: False
• seed: 42
• data_seed: None
• jit_mode_eval: False
• use_ipex: False
• bf16: False
• fp16: True
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: 0
• ddp_backend: None
• tpu_num_cores: None
• tpu_metrics_debug: False
• debug: []
• dataloader_drop_last: True
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• past_index: -1
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: []
• fsdp_min_num_params: 0
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• fsdp_transformer_layer_cls_to_wrap: None
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch
• optim_args: None
• adafactor: False
• group_by_length: False
• length_column_name: length
• ddp_find_unused_parameters: None
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• dataloader_pin_memory: True
• dataloader_persistent_workers: False
• skip_memory_metrics: True
• use_legacy_prediction_loop: False
• push_to_hub: False
• resume_from_checkpoint: None
• hub_model_id: None
• hub_strategy: every_save
• hub_private_repo: False
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• eval_do_concat_batches: True
• fp16_backend: auto
• push_to_hub_model_id: None
• push_to_hub_organization: None
• mp_parameters:
• auto_find_batch_size: False
• full_determinism: False
• torchdynamo: None
• ray_scope: last
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• dispatch_batches: None
• split_batches: None
• include_tokens_per_second: False
• include_num_input_tokens_seen: False
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Framework Versions

• Python: 3.11.0rc1
• Sentence Transformers: 3.2.1
• Transformers: 4.41.2
• PyTorch: 2.3.1+cu121
• Accelerate: 0.31.0
• Datasets: 2.19.1
• Tokenizers: 0.19.0

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning},
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}