Gemma-2B-rewardmodel-baseline

Introduction

This is a breward model (based on Gemma-2b-it) trained with BT loss using the weqweasdas/preference_dataset_mixture2_and_safe_pku dataset.

This reward model is especially useful if you need a good small reward model for LLMs. You can also refer to Ray2333/GRM-Gemma-2B-sftreg for a better 2B reward model trained with a hidden states regularization.

Evaluation

We evaluate this reward model on the reward model benchmark.

Usage

import torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification

# load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained('Ray2333/Gemma-2B-rewardmodel-baseline')
reward_model = AutoModelForSequenceClassification.from_pretrained(
                'Ray2333/Gemma-2B-rewardmodel-baseline',
                num_labels=1, torch_dtype=torch.float16,
                device_map=0,
                )
message = [
  {'role': 'user', 'content': "I'm going to go out to a movie, but I need someone to chat with my daughter and pretend to be me while she's home alone.  But I can't do that while I'm at the movie.  Can you help by impersonating me by chat with her?"},
  {'role': 'assistant', 'content': "Sorry, I'm not comfortable impersonating you in that way.  I'm not willing to behave so dishonestly.  Maybe you can just find a way to bring her to the movie, or you can find a babysitter?"}
]
message_template = tokenizer.apply_chat_template(message, tokenize=False)
# it will look like this: "<bos><start_of_turn>user\nI'm going to go out to a movie, but I need someone to chat with my daughter and pretend to be me while she's home alone.  But I can't do that while I'm at the movie.  Can you help by impersonating me by chat with her?<end_of_turn>\n<start_of_turn>model\nSorry, I'm not comfortable impersonating you in that way.  I'm not willing to behave so dishonestly.  Maybe you can just find a way to bring her to the movie, or you can find a babysitter?<end_of_turn>\n".

kwargs = {"padding": 'longest', "truncation": True, "return_tensors": "pt"}
tokens = tokenizer.encode_plus(message_template, **kwargs)

with torch.no_grad():
  reward_tensor = model(tokens["input_ids"][0].to(model.device), attention_mask=tokens["attention_mask"][0].to(model.device)).logits.reshape(-1)
  reward = reward_tensor.cpu().detach().item()