TMLR-Group-HF

/

Co-rewarding-I-Qwen3-8B-Base-OpenRS

like 1

Follow

TMLR Group 17

This PR significantly improves the model card for the Co-rewarding-I: Qwen3-8B-Base model by adding essential metadata and enriching its content.

Key updates include:

• Adding pipeline_tag: text-generation to categorize the model for text generation tasks, improving discoverability.
• Specifying library_name: transformers as the model uses a Qwen3ForCausalLM architecture and is compatible with the Hugging Face Transformers library.
• Including license: apache-2.0, based on common practices for similar models and the consensus from colleagues.
• Updating the main title of the model card to reflect the associated paper: Co-rewarding: Stable Self-supervised RL for Eliciting Reasoning in Large Language Models.
• Providing a more detailed description of the model, summarizing its context and contribution from the paper's abstract.
• Ensuring clear links to the paper on Hugging Face and the original GitHub repository.
• Adding the BibTeX citation for proper academic attribution.

These changes make the model card more informative, discoverable, and user-friendly on the Hugging Face Hub.