README.md

---
language:
- en
license: llama3
tags:
- text-classification
datasets:
- openbmb/UltraFeedback
- nvidia/HelpSteer
- Anthropic/hh-rlhf
- PKU-Alignment/PKU-SafeRLHF
- NCSOFT/offsetbias
base_model:
- sfairXC/FsfairX-LLaMA3-RM-v0.1
- meta-llama/Meta-Llama-3-8B-Instruct
---

# Model Card for Llama-3-OffsetBias-RM-8B

**Llama-3-OffsetBias-RM-8B** is a *reward model* trained on OffsetBias dataset. It is trained to be more robust on various evaluation *biases* commonly found in evaluation models. The model is introduced in paper **OffsetBias: Leveraging Debiased Data for Tuning Evaluators**.

## Model Details

### Model Description

**Llama-3-OffsetBias-RM-8B** uses [sfairXC/FsfairX-LLaMA3-RM-v0.1](https://huggingface.co/sfairXC/FsfairX-LLaMA3-RM-v0.1) as base model, which is built with Meta Llama 3. An intermediate reward model is trained from from [Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) using a subset of dataset used in training of *FsfairX-LLaMA3-RM* model, combined with *NCSOFT/offsetbias* dataset. The intermediate model is then merged with *FsfairX-LLaMA3-RM* model to create **Llama-3-OffsetBias-RM-8B**.

- **Developed by:** NC Research
- **Language(s) (NLP):** English
- **License:** llama3
- **Finetuned from model:** [sfairXC/FsfairX-LLaMA3-RM-v0.1](https://huggingface.co/sfairXC/FsfairX-LLaMA3-RM-v0.1)

### Model Sources

- 💻 **Repository:** [https://github.com/ncsoft/offsetbias](https://github.com/ncsoft/offsetbias)
- 📜 **Paper:** [OffsetBias: Leveraging Debiased Data for Tuning Evaluators](https://arxiv.org/abs/2407.06551)
- 🤗 **Dataset:** [https://huggingface.co/datasets/NCSOFT/offsetbias](https://huggingface.co/datasets/NCSOFT/offsetbias)

## Uses

### Direct Use

```python
  from transformers import AutoTokenizer, pipeline

  model_name = "NCSOFT/Llama-3-OffsetBias-RM-8B"
  rm_tokenizer = AutoTokenizer.from_pretrained(model_name)
  rm_pipe = pipeline(
      "sentiment-analysis",
      model=model_name,
      device="auto",
      tokenizer=rm_tokenizer,
      model_kwargs={"torch_dtype": torch.bfloat16}
  )

  pipe_kwargs = {
      "return_all_scores": True,
      "function_to_apply": "none",
      "batch_size": 1
  }

  chat = [
   {"role": "user", "content": "Hello, how are you?"},
   {"role": "assistant", "content": "I'm doing great. How can I help you today?"},
   {"role": "user", "content": "I'd like to show off how chat templating works!"},
  ]

  test_texts = [tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=False).replace(tokenizer.bos_token, "")]
  pipe_outputs = rm_pipe(test_texts, **pipe_kwargs)
  rewards = [output[0]["score"] for output in pipe_outputs]
```

## Evaluation

### RewardBench Result
| Metric       | Score  |
|--------------|--------|
| Chat         | 97.21  |
| Chat Hard    | 80.70  |
| Safety       | 89.01  |
| Reasoning    | 90.60  |

### EvalBiasBench Result

| Metric                | Score |
|-----------------------|-------|
| Length                | 82.4  |
| Concreteness          | 92.9  |
| Empty Reference       | 46.2  |
| Content Continuation  | 100.0 |
| Nested Instruction    | 83.3  |
| Familiar Knowledge    | 58.3  |

## Citation

```bibtex
@misc{park2024offsetbias,
      title={OffsetBias: Leveraging Debiased Data for Tuning Evaluators},
      author={Junsoo Park and Seungyeon Jwa and Meiying Ren and Daeyoung Kim and Sanghyuk Choi},
      year={2024},
      eprint={2407.06551},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}
```