---
license: apache-2.0
datasets:
- scientific_papers
metrics:
- bertscore
- rouge
tags:
- text-generation-inference
- rlhf
- PPO
language:
- en
---

This model is fintuned using PPO based NLPO RL algorithm, on ccdv/arxiv-summarization dataset. The base model is flan-t5-base model.