You can deploy this model and start earning money today!

GenuineWWD/
SCS_data

Views

No views yet

[NeurIPS 2025] Enhancing the Outcome Reward-based RL Training of MLLMs with Self-Consistency Sampling

A simple, general sampling method for RLVR with multi-choice dataset to solve unfaithful reasoning phenomenon!

SCS Resouces

📖 Paper | 🤗 Dataset | 💻 Code

🔔News

🔥[2025-11-9] Release the eval codes! 🚀 🔥[2025-10-13] Release the dataset the codes! 🚀 🔥[2025-9-17] Our SCS paper is accepted by NeurIPS 2025! 🚀

To-do

Release the eval codes… See the full description on the dataset page: https://huggingface.co/datasets/GenuineWWD/SCS_data.