README.md

metadata

tags:
  - merge
  - mergekit
  - lazymergekit
  - BlackBeenie/Neos-Llama-3.1-8B
  - Solshine/Meta-Llama-3.1-8B-Instruct-Python-Coder
  - Solshine/reflection-llama-3.1-8B
base_model:
  - BlackBeenie/Neos-Llama-3.1-8B
  - Solshine/Meta-Llama-3.1-8B-Instruct-Python-Coder
  - Solshine/reflection-llama-3.1-8B
model-index:
  - name: Bloslain-8B-v0.2
    results:
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: IFEval (0-Shot)
          type: HuggingFaceH4/ifeval
          args:
            num_few_shot: 0
        metrics:
          - type: inst_level_strict_acc and prompt_level_strict_acc
            value: 50.23
            name: strict accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=BlackBeenie/Bloslain-8B-v0.2
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: BBH (3-Shot)
          type: BBH
          args:
            num_few_shot: 3
        metrics:
          - type: acc_norm
            value: 30.66
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=BlackBeenie/Bloslain-8B-v0.2
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MATH Lvl 5 (4-Shot)
          type: hendrycks/competition_math
          args:
            num_few_shot: 4
        metrics:
          - type: exact_match
            value: 14.5
            name: exact match
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=BlackBeenie/Bloslain-8B-v0.2
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: GPQA (0-shot)
          type: Idavidrein/gpqa
          args:
            num_few_shot: 0
        metrics:
          - type: acc_norm
            value: 7.49
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=BlackBeenie/Bloslain-8B-v0.2
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MuSR (0-shot)
          type: TAUR-Lab/MuSR
          args:
            num_few_shot: 0
        metrics:
          - type: acc_norm
            value: 10.45
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=BlackBeenie/Bloslain-8B-v0.2
          name: Open LLM Leaderboard
      - task:
          type: text-generation
          name: Text Generation
        dataset:
          name: MMLU-PRO (5-shot)
          type: TIGER-Lab/MMLU-Pro
          config: main
          split: test
          args:
            num_few_shot: 5
        metrics:
          - type: acc
            value: 29.48
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=BlackBeenie/Bloslain-8B-v0.2
          name: Open LLM Leaderboard

Bloslain-8B-v0.2

Bloslain-8B-v0.2 is a merge of the following models using LazyMergekit:

• BlackBeenie/Neos-Llama-3.1-8B
• Solshine/Meta-Llama-3.1-8B-Instruct-Python-Coder
• Solshine/reflection-llama-3.1-8B

🧩 Configuration

models:
  - model: NousResearch/Meta-Llama-3.1-8B-Instruct
    # No parameters necessary for base model
  - model: BlackBeenie/Neos-Llama-3.1-8B
    parameters:
      density: 0.53
      weight: 0.4
  - model: Solshine/Meta-Llama-3.1-8B-Instruct-Python-Coder
    parameters:
      density: 0.53
      weight: 0.3
  - model: Solshine/reflection-llama-3.1-8B
    parameters:
      density: 0.53
      weight: 0.3
merge_method: dare_ties
base_model: NousResearch/Meta-Llama-3.1-8B-Instruct
parameters:
  int8_mask: true
dtype: bfloat16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "BlackBeenie/Bloslain-8B-v0.2"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	23.80
IFEval (0-Shot)	50.23
BBH (3-Shot)	30.66
MATH Lvl 5 (4-Shot)	14.50
GPQA (0-shot)	7.49
MuSR (0-shot)	10.45
MMLU-PRO (5-shot)	29.48