README.md

metadata

language:
  - en
license: apache-2.0
tags:
  - chat
pipeline_tag: text-generation
model-index:
  - name: Qwen2-7B-Instruct-abliterated
    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: 58.37
            name: strict accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=natong19/Qwen2-7B-Instruct-abliterated
          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: 37.75
            name: normalized accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=natong19/Qwen2-7B-Instruct-abliterated
          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: 10.27
            name: exact match
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=natong19/Qwen2-7B-Instruct-abliterated
          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: 6.82
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=natong19/Qwen2-7B-Instruct-abliterated
          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: 8.93
            name: acc_norm
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=natong19/Qwen2-7B-Instruct-abliterated
          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: 31.58
            name: accuracy
        source:
          url: >-
            https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=natong19/Qwen2-7B-Instruct-abliterated
          name: Open LLM Leaderboard

Qwen2-7B-Instruct-abliterated

Introduction

Abliterated version of Qwen2-7B-Instruct using failspy's notebook. The model's strongest refusal directions have been ablated via weight orthogonalization, but the model may still refuse your request, misunderstand your intent, or provide unsolicited advice regarding ethics or safety.

Quickstart

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "natong19/Qwen2-7B-Instruct-abliterated"
device = "cuda" # the device to load the model onto

model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained(model_id)

prompt = "Give me a short introduction to large language model."
messages = [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)
model_inputs = tokenizer([text], return_tensors="pt").to(device)

generated_ids = model.generate(
    model_inputs.input_ids,
    max_new_tokens=256
)
generated_ids = [
    output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]

response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
print(response)

Evaluation

Evaluation framework: lm-evaluation-harness 0.4.2

Datasets	Qwen2-7B-Instruct	Qwen2-7B-Instruct-abliterated
ARC (25-shot)	62.5	62.5
GSM8K (5-shot)	73.0	72.2
HellaSwag (10-shot)	81.8	81.7
MMLU (5-shot)	70.7	70.5
TruthfulQA (0-shot)	57.3	55.0
Winogrande (5-shot)	76.2	77.4

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	25.62
IFEval (0-Shot)	58.37
BBH (3-Shot)	37.75
MATH Lvl 5 (4-Shot)	10.27
GPQA (0-shot)	6.82
MuSR (0-shot)	8.93
MMLU-PRO (5-shot)	31.58