QRM-Llama3-8B

Quantile Regression for Distributional Reward Models in RLHF

(This is an old version. The new one trained on the decontaminated version of the Skywork dataset is nicolinho/QRM-Llama3.1-8B-v2)

• Author: Nicolai Dorka

• Tech Report: https://arxiv.org/abs/2409.10164

• Code Repository: https://github.com/Nicolinho/QRM

• Method Overview: QRM generates a distribution over rewards by aggregating individual distributions over attribute scores like helpfulness and harmlessness.

  <p align="left"> <img width="800" alt="image" src="https://github.com/Nicolinho/QRM/blob/main/assets/method_vis.png?raw=true"> </p>

Demo Code

import torch
from transformers import AutoModelForSequenceClassification, AutoTokenizer
device = "cuda"
path = "nicolinho/QRM-Llama3-8B"
model = AutoModelForSequenceClassification.from_pretrained(path, device_map=device, 
                               trust_remote_code=True, torch_dtype=torch.bfloat16)
tokenizer = AutoTokenizer.from_pretrained(path, use_fast=True)
# We load a random sample from the validation set of the HelpSteer dataset
prompt = 'Does pineapple belong on a Pizza?'
response = "There are different opinions on this. Some people like pineapple on a Pizza while others condemn this."
messages = [{"role": "user", "content": prompt},
           {"role": "assistant", "content": response}]
input_ids = tokenizer.apply_chat_template(messages, return_tensors="pt").to(device)
with torch.no_grad():
   output = model(input_ids)
   # Expectation of the reward distribution
   reward = output.score.cpu().float() 
   # Quantile estimates for the quantiles 0.05, 0.1, ..., 0.9, 0.95 representing the distribution over rewards
   reward_quantiles = output.reward_quantiles.cpu().float()

# The attributes of the 19 reward objectives
attributes = ['helpsteer-helpfulness','helpsteer-correctness','helpsteer-coherence',
   'helpsteer-complexity','helpsteer-verbosity','ultrafeedback-overall_score',
   'ultrafeedback-instruction_following', 'ultrafeedback-truthfulness',
   'ultrafeedback-honesty','ultrafeedback-helpfulness','beavertails-is_safe',
   'prometheus-score','argilla-overall_quality','argilla-judge_lm','code-complexity',
   'code-style','code-explanation','code-instruction-following','code-readability']

Citation

If you find this work useful for your research, please consider citing:

@article{dorka2024quantile,
  title={Quantile Regression for Distributional Reward Models in RLHF},
  author={Dorka, Nicolai},
  journal={arXiv preprint arXiv:2409.10164},
  year={2024}
}