This repository contains the data of the paper "Agentic Reward Modeling: Integrating Human Preferences with Verifiable Correctness Signals for Reliable Reward Systems"
Paper:
https://arxiv.org/abs/2502.19328
GitHub:
https://github.com/THU-KEG/Agentic-Reward-Modelingthe samples are formatted as follows:
{
"id": // unique identifier of the sample,
"source": // source… See the full description on the dataset page:
https://huggingface.co/datasets/THU-KEG/IFBench.
