File size: 1,625 Bytes
a888002 7556380 a888002 7556380 a888002 7556380 a888002 7556380 a888002 7556380 a888002 7556380 a888002 7556380 a888002 7556380 a888002 7556380 a888002 7556380 a888002 7556380 a888002 7556380 a888002 7556380 a888002 7556380 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 |
---
library_name: transformers
license: mit
datasets:
- hendrydong/preference_700K
base_model:
- RLHFlow/LLaMA3-SFT-v2
pipeline_tag: text-classification
---
# rlhflow-llama-3-sft-segment Model Card
- **Paper:** [Segmenting Text and Learning Their Rewards for Improved RLHF in Language Model
](https://arxiv.org/abs/2501.02790)
- **Model:** [yyqoni/rlhflow-llama-3-sft-8b-v2-segment-rm-700k](https://huggingface.co/yyqoni/rlhflow-llama-3-sft-8b-v2-segment-rm-700k)
## Method
The segment reward model assigns rewards to semantically meaningful text segments, segmented dynamically with an entropy-based threshold. It is trained on binary preference labels from human feedback, optimizing a Bradley-Terry loss function that aggregates segment rewards using the average function.
## Architecture
<div align=center>
</div>
## Training
The phi-instruct-segment model is fine-tuned from **RLHFlow/LLaMA3-SFT-v2** on the **hendrydong/preference_700K dataset**.
## Citation
If you find this model or our research useful, please consider citing our paper:
```bibtex
@misc{yin2025segmentingtextlearningrewards,
title={Segmenting Text and Learning Their Rewards for Improved RLHF in Language Model},
author={Yueqin Yin and Shentao Yang and Yujia Xie and Ziyi Yang and Yuting Sun and Hany Awadalla and Weizhu Chen and Mingyuan Zhou},
year={2025},
eprint={2501.02790},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2501.02790},
}
``` |