metadata
language:
- en
license: apache-2.0
tags:
- gpt2
- dpo
- trl
datasets:
- HuggingFaceH4/ultrachat_200k
- Intel/orca_dpo_pairs
pipeline_tag: text-generation
model-index:
- name: chat_gpt2_dpo
results:
- task:
type: text-generation
name: Text Generation
dataset:
name: AI2 Reasoning Challenge (25-Shot)
type: ai2_arc
config: ARC-Challenge
split: test
args:
num_few_shot: 25
metrics:
- type: acc_norm
value: 23.98
name: normalized accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Sharathhebbar24/chat_gpt2_dpo
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: HellaSwag (10-Shot)
type: hellaswag
split: validation
args:
num_few_shot: 10
metrics:
- type: acc_norm
value: 31.22
name: normalized accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Sharathhebbar24/chat_gpt2_dpo
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: MMLU (5-Shot)
type: cais/mmlu
config: all
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 24.95
name: accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Sharathhebbar24/chat_gpt2_dpo
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: TruthfulQA (0-shot)
type: truthful_qa
config: multiple_choice
split: validation
args:
num_few_shot: 0
metrics:
- type: mc2
value: 41.26
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Sharathhebbar24/chat_gpt2_dpo
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: Winogrande (5-shot)
type: winogrande
config: winogrande_xl
split: validation
args:
num_few_shot: 5
metrics:
- type: acc
value: 49.96
name: accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Sharathhebbar24/chat_gpt2_dpo
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: GSM8k (5-shot)
type: gsm8k
config: main
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 0
name: accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=Sharathhebbar24/chat_gpt2_dpo
name: Open LLM Leaderboard
This model is a finetuned version of Sharathhebbar24/chat_gpt2 using Intel/orca_dpo_pairs on DPO
Model description
GPT-2 is a transformers model pre-trained on a very large corpus of English data in a self-supervised fashion. This means it was pre-trained on the raw texts only, with no humans labeling them in any way (which is why it can use lots of publicly available data) with an automatic process to generate inputs and labels from those texts. More precisely, it was trained to guess the next word in sentences.
More precisely, inputs are sequences of continuous text of a certain length and the targets are the same sequence,
shifting one token (word or piece of word) to the right. The model uses a masking mechanism to make sure the
predictions for the token i only use the inputs from 1 to i but not the future tokens.
This way, the model learns an inner representation of the English language that can then be used to extract features useful for downstream tasks. The model is best at what it was trained for, however, which is generating texts from a prompt.
To use this model
>>> from transformers import AutoTokenizer, AutoModelForCausalLM
>>> model_name = "Sharathhebbar24/chat_gpt2_dpo"
>>> model = AutoModelForCausalLM.from_pretrained(model_name)
>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
>>> def generate_text(prompt):
>>> inputs = tokenizer.encode(prompt, return_tensors='pt')
>>> outputs = model.generate(inputs, max_length=64, pad_token_id=tokenizer.eos_token_id)
>>> generated = tokenizer.decode(outputs[0], skip_special_tokens=True)
>>> return generated[:generated.rfind(".")+1]
>>> prompt = """
>>> user: what are you?
>>> assistant: I am a Chatbot intended to give a python program
>>> user: hmm, can you write a python program to print Hii Heloo
>>> assistant: Sure Here is a python code.\n print("Hii Heloo")
>>> user: Can you write a Linear search program in python
>>> """
>>> res = generate_text(prompt)
>>> res
Benchmark / Evaluation
| Model | Average | ARC | HellaSwag | MMLU | TruthfulQA | Winogrande | GSM8k |
|---|---|---|---|---|---|---|---|
| Sharathhebbar24/chat_gpt2_dpo | 28.56 | 23.98 | 31.22 | 24.95 | 41.26 | 49.96 | 0 |
{
"all": {
"acc": 0.24915779048270345,
"acc_stderr": 0.030509906389610868,
"acc_norm": 0.25041231816215265,
"acc_norm_stderr": 0.03132600249114931,
"mc1": 0.2521419828641371,
"mc1_stderr": 0.015201522246299965,
"mc2": 0.41257163824244014,
"mc2_stderr": 0.015127188811834062
},
"harness|arc:challenge|25": {
"acc": 0.18686006825938567,
"acc_stderr": 0.011391015649694391,
"acc_norm": 0.23976109215017063,
"acc_norm_stderr": 0.012476304127453954
},
"harness|hellaswag|10": {
"acc": 0.28978291177056364,
"acc_stderr": 0.004527343651130803,
"acc_norm": 0.3121888070105557,
"acc_norm_stderr": 0.0046243936909668975
},
"harness|hendrycksTest-abstract_algebra|5": {
"acc": 0.22,
"acc_stderr": 0.04163331998932268,
"acc_norm": 0.22,
"acc_norm_stderr": 0.04163331998932268
},
"harness|hendrycksTest-anatomy|5": {
"acc": 0.3037037037037037,
"acc_stderr": 0.039725528847851375,
"acc_norm": 0.3037037037037037,
"acc_norm_stderr": 0.039725528847851375
},
"harness|hendrycksTest-astronomy|5": {
"acc": 0.17763157894736842,
"acc_stderr": 0.031103182383123398,
"acc_norm": 0.17763157894736842,
"acc_norm_stderr": 0.031103182383123398
},
"harness|hendrycksTest-business_ethics|5": {
"acc": 0.26,
"acc_stderr": 0.0440844002276808,
"acc_norm": 0.26,
"acc_norm_stderr": 0.0440844002276808
},
"harness|hendrycksTest-clinical_knowledge|5": {
"acc": 0.23018867924528302,
"acc_stderr": 0.025907897122408173,
"acc_norm": 0.23018867924528302,
"acc_norm_stderr": 0.025907897122408173
},
"harness|hendrycksTest-college_biology|5": {
"acc": 0.2569444444444444,
"acc_stderr": 0.03653946969442099,
"acc_norm": 0.2569444444444444,
"acc_norm_stderr": 0.03653946969442099
},
"harness|hendrycksTest-college_chemistry|5": {
"acc": 0.19,
"acc_stderr": 0.039427724440366234,
"acc_norm": 0.19,
"acc_norm_stderr": 0.039427724440366234
},
"harness|hendrycksTest-college_computer_science|5": {
"acc": 0.24,
"acc_stderr": 0.04292346959909283,
"acc_norm": 0.24,
"acc_norm_stderr": 0.04292346959909283
},
"harness|hendrycksTest-college_mathematics|5": {
"acc": 0.29,
"acc_stderr": 0.04560480215720684,
"acc_norm": 0.29,
"acc_norm_stderr": 0.04560480215720684
},
"harness|hendrycksTest-college_medicine|5": {
"acc": 0.2543352601156069,
"acc_stderr": 0.0332055644308557,
"acc_norm": 0.2543352601156069,
"acc_norm_stderr": 0.0332055644308557
},
"harness|hendrycksTest-college_physics|5": {
"acc": 0.21568627450980393,
"acc_stderr": 0.04092563958237654,
"acc_norm": 0.21568627450980393,
"acc_norm_stderr": 0.04092563958237654
},
"harness|hendrycksTest-computer_security|5": {
"acc": 0.34,
"acc_stderr": 0.04760952285695236,
"acc_norm": 0.34,
"acc_norm_stderr": 0.04760952285695236
},
"harness|hendrycksTest-conceptual_physics|5": {
"acc": 0.26382978723404255,
"acc_stderr": 0.028809989854102973,
"acc_norm": 0.26382978723404255,
"acc_norm_stderr": 0.028809989854102973
},
"harness|hendrycksTest-econometrics|5": {
"acc": 0.24561403508771928,
"acc_stderr": 0.04049339297748142,
"acc_norm": 0.24561403508771928,
"acc_norm_stderr": 0.04049339297748142
},
"harness|hendrycksTest-electrical_engineering|5": {
"acc": 0.2413793103448276,
"acc_stderr": 0.03565998174135302,
"acc_norm": 0.2413793103448276,
"acc_norm_stderr": 0.03565998174135302
},
"harness|hendrycksTest-elementary_mathematics|5": {
"acc": 0.24074074074074073,
"acc_stderr": 0.02201908001221789,
"acc_norm": 0.24074074074074073,
"acc_norm_stderr": 0.02201908001221789
},
"harness|hendrycksTest-formal_logic|5": {
"acc": 0.1349206349206349,
"acc_stderr": 0.030557101589417515,
"acc_norm": 0.1349206349206349,
"acc_norm_stderr": 0.030557101589417515
},
"harness|hendrycksTest-global_facts|5": {
"acc": 0.16,
"acc_stderr": 0.03684529491774708,
"acc_norm": 0.16,
"acc_norm_stderr": 0.03684529491774708
},
"harness|hendrycksTest-high_school_biology|5": {
"acc": 0.1774193548387097,
"acc_stderr": 0.02173254068932927,
"acc_norm": 0.1774193548387097,
"acc_norm_stderr": 0.02173254068932927
},
"harness|hendrycksTest-high_school_chemistry|5": {
"acc": 0.24630541871921183,
"acc_stderr": 0.030315099285617736,
"acc_norm": 0.24630541871921183,
"acc_norm_stderr": 0.030315099285617736
},
"harness|hendrycksTest-high_school_computer_science|5": {
"acc": 0.28,
"acc_stderr": 0.04512608598542126,
"acc_norm": 0.28,
"acc_norm_stderr": 0.04512608598542126
},
"harness|hendrycksTest-high_school_european_history|5": {
"acc": 0.21818181818181817,
"acc_stderr": 0.03225078108306289,
"acc_norm": 0.21818181818181817,
"acc_norm_stderr": 0.03225078108306289
},
"harness|hendrycksTest-high_school_geography|5": {
"acc": 0.3282828282828283,
"acc_stderr": 0.03345678422756776,
"acc_norm": 0.3282828282828283,
"acc_norm_stderr": 0.03345678422756776
},
"harness|hendrycksTest-high_school_government_and_politics|5": {
"acc": 0.37305699481865284,
"acc_stderr": 0.03490205592048573,
"acc_norm": 0.37305699481865284,
"acc_norm_stderr": 0.03490205592048573
},
"harness|hendrycksTest-high_school_macroeconomics|5": {
"acc": 0.26666666666666666,
"acc_stderr": 0.02242127361292371,
"acc_norm": 0.26666666666666666,
"acc_norm_stderr": 0.02242127361292371
},
"harness|hendrycksTest-high_school_mathematics|5": {
"acc": 0.21481481481481482,
"acc_stderr": 0.025040443877000683,
"acc_norm": 0.21481481481481482,
"acc_norm_stderr": 0.025040443877000683
},
"harness|hendrycksTest-high_school_microeconomics|5": {
"acc": 0.22268907563025211,
"acc_stderr": 0.027025433498882364,
"acc_norm": 0.22268907563025211,
"acc_norm_stderr": 0.027025433498882364
},
"harness|hendrycksTest-high_school_physics|5": {
"acc": 0.23178807947019867,
"acc_stderr": 0.034454062719870546,
"acc_norm": 0.23178807947019867,
"acc_norm_stderr": 0.034454062719870546
},
"harness|hendrycksTest-high_school_psychology|5": {
"acc": 0.3302752293577982,
"acc_stderr": 0.02016446633634298,
"acc_norm": 0.3302752293577982,
"acc_norm_stderr": 0.02016446633634298
},
"harness|hendrycksTest-high_school_statistics|5": {
"acc": 0.19444444444444445,
"acc_stderr": 0.026991454502036733,
"acc_norm": 0.19444444444444445,
"acc_norm_stderr": 0.026991454502036733
},
"harness|hendrycksTest-high_school_us_history|5": {
"acc": 0.25,
"acc_stderr": 0.03039153369274154,
"acc_norm": 0.25,
"acc_norm_stderr": 0.03039153369274154
},
"harness|hendrycksTest-high_school_world_history|5": {
"acc": 0.26582278481012656,
"acc_stderr": 0.028756799629658342,
"acc_norm": 0.26582278481012656,
"acc_norm_stderr": 0.028756799629658342
},
"harness|hendrycksTest-human_aging|5": {
"acc": 0.17937219730941703,
"acc_stderr": 0.0257498195691928,
"acc_norm": 0.17937219730941703,
"acc_norm_stderr": 0.0257498195691928
},
"harness|hendrycksTest-human_sexuality|5": {
"acc": 0.2366412213740458,
"acc_stderr": 0.037276735755969174,
"acc_norm": 0.2366412213740458,
"acc_norm_stderr": 0.037276735755969174
},
"harness|hendrycksTest-international_law|5": {
"acc": 0.35537190082644626,
"acc_stderr": 0.04369236326573981,
"acc_norm": 0.35537190082644626,
"acc_norm_stderr": 0.04369236326573981
},
"harness|hendrycksTest-jurisprudence|5": {
"acc": 0.25925925925925924,
"acc_stderr": 0.042365112580946336,
"acc_norm": 0.25925925925925924,
"acc_norm_stderr": 0.042365112580946336
},
"harness|hendrycksTest-logical_fallacies|5": {
"acc": 0.2822085889570552,
"acc_stderr": 0.03536117886664742,
"acc_norm": 0.2822085889570552,
"acc_norm_stderr": 0.03536117886664742
},
"harness|hendrycksTest-machine_learning|5": {
"acc": 0.32142857142857145,
"acc_stderr": 0.04432804055291519,
"acc_norm": 0.32142857142857145,
"acc_norm_stderr": 0.04432804055291519
},
"harness|hendrycksTest-management|5": {
"acc": 0.1941747572815534,
"acc_stderr": 0.03916667762822585,
"acc_norm": 0.1941747572815534,
"acc_norm_stderr": 0.03916667762822585
},
"harness|hendrycksTest-marketing|5": {
"acc": 0.2905982905982906,
"acc_stderr": 0.02974504857267404,
"acc_norm": 0.2905982905982906,
"acc_norm_stderr": 0.02974504857267404
},
"harness|hendrycksTest-medical_genetics|5": {
"acc": 0.25,
"acc_stderr": 0.04351941398892446,
"acc_norm": 0.25,
"acc_norm_stderr": 0.04351941398892446
},
"harness|hendrycksTest-miscellaneous|5": {
"acc": 0.23627075351213284,
"acc_stderr": 0.015190473717037497,
"acc_norm": 0.23627075351213284,
"acc_norm_stderr": 0.015190473717037497
},
"harness|hendrycksTest-moral_disputes|5": {
"acc": 0.24566473988439305,
"acc_stderr": 0.02317629820399201,
"acc_norm": 0.24566473988439305,
"acc_norm_stderr": 0.02317629820399201
},
"harness|hendrycksTest-moral_scenarios|5": {
"acc": 0.24581005586592178,
"acc_stderr": 0.014400296429225587,
"acc_norm": 0.24581005586592178,
"acc_norm_stderr": 0.014400296429225587
},
"harness|hendrycksTest-nutrition|5": {
"acc": 0.25163398692810457,
"acc_stderr": 0.024848018263875195,
"acc_norm": 0.25163398692810457,
"acc_norm_stderr": 0.024848018263875195
},
"harness|hendrycksTest-philosophy|5": {
"acc": 0.18006430868167203,
"acc_stderr": 0.021823422857744953,
"acc_norm": 0.18006430868167203,
"acc_norm_stderr": 0.021823422857744953
},
"harness|hendrycksTest-prehistory|5": {
"acc": 0.25617283950617287,
"acc_stderr": 0.024288533637726095,
"acc_norm": 0.25617283950617287,
"acc_norm_stderr": 0.024288533637726095
},
"harness|hendrycksTest-professional_accounting|5": {
"acc": 0.2801418439716312,
"acc_stderr": 0.02678917235114023,
"acc_norm": 0.2801418439716312,
"acc_norm_stderr": 0.02678917235114023
},
"harness|hendrycksTest-professional_law|5": {
"acc": 0.24837027379400262,
"acc_stderr": 0.011035212598034503,
"acc_norm": 0.24837027379400262,
"acc_norm_stderr": 0.011035212598034503
},
"harness|hendrycksTest-professional_medicine|5": {
"acc": 0.3125,
"acc_stderr": 0.02815637344037142,
"acc_norm": 0.3125,
"acc_norm_stderr": 0.02815637344037142
},
"harness|hendrycksTest-professional_psychology|5": {
"acc": 0.25,
"acc_stderr": 0.01751781884501444,
"acc_norm": 0.25,
"acc_norm_stderr": 0.01751781884501444
},
"harness|hendrycksTest-public_relations|5": {
"acc": 0.18181818181818182,
"acc_stderr": 0.03694284335337801,
"acc_norm": 0.18181818181818182,
"acc_norm_stderr": 0.03694284335337801
},
"harness|hendrycksTest-security_studies|5": {
"acc": 0.31020408163265306,
"acc_stderr": 0.029613459872484378,
"acc_norm": 0.31020408163265306,
"acc_norm_stderr": 0.029613459872484378
},
"harness|hendrycksTest-sociology|5": {
"acc": 0.24875621890547264,
"acc_stderr": 0.030567675938916707,
"acc_norm": 0.24875621890547264,
"acc_norm_stderr": 0.030567675938916707
},
"harness|hendrycksTest-us_foreign_policy|5": {
"acc": 0.25,
"acc_stderr": 0.04351941398892446,
"acc_norm": 0.25,
"acc_norm_stderr": 0.04351941398892446
},
"harness|hendrycksTest-virology|5": {
"acc": 0.19879518072289157,
"acc_stderr": 0.03106939026078942,
"acc_norm": 0.19879518072289157,
"acc_norm_stderr": 0.03106939026078942
},
"harness|hendrycksTest-world_religions|5": {
"acc": 0.29239766081871343,
"acc_stderr": 0.034886477134579215,
"acc_norm": 0.29239766081871343,
"acc_norm_stderr": 0.034886477134579215
},
"harness|truthfulqa:mc|0": {
"mc1": 0.2521419828641371,
"mc1_stderr": 0.015201522246299965,
"mc2": 0.41257163824244014,
"mc2_stderr": 0.015127188811834062
},
"harness|winogrande|5": {
"acc": 0.4996053670086819,
"acc_stderr": 0.014052481306049512
},
"harness|gsm8k|5": {
"acc": 0.0,
"acc_stderr": 0.0
}
}
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
| Metric | Value |
|---|---|
| Avg. | 28.56 |
| AI2 Reasoning Challenge (25-Shot) | 23.98 |
| HellaSwag (10-Shot) | 31.22 |
| MMLU (5-Shot) | 24.95 |
| TruthfulQA (0-shot) | 41.26 |
| Winogrande (5-shot) | 49.96 |
| GSM8k (5-shot) | 0.00 |