tuhink/hacking-rewards-sft-llama
Text Generation • 3B • Updated • 16
Tuhin Kundu
tuhink
AI & ML interests
Vision, language, machine learning systems
Recent Activity
upvoted an article 5 days ago
Turning Database Grading into RL Rewards: Post-Training Agents with ThinkingBox reacted to sergiopaniego's post with 👍 5 days ago
ThinkingBox from @microsoft is now available as an OpenEnv env (cc @tuhink🤗 )!
> ThinkingBox is a sandbox for testing agents on business workflows. it simulates a customer, gives the agent MCP tools over a real database, and at the end checks what changed in that database instead of trusting the agent's last message
> ThinkingBox-Bench is the benchmark built on it: 507 tasks across retail, insurance, travel, banking and consulting
> the OpenEnv env runs each task as an episode in its own isolated backend and returns a pass/fail reward from those checks
https://huggingface.co/blog/microsoft/thinkingbox published an article 5 days ago
Turning Database Grading into RL Rewards: Post-Training Agents with ThinkingBoxOrganizations
tuhink/hacking-rewards-general-train
Text Classification • 0.4B • Updated • 15
tuhink/hacking-rewards-coherence-train
Text Classification • 0.4B • Updated • 9
tuhink/hacking-rewards-harmless-train
Text Classification • 0.4B • Updated • 8
tuhink/hacking-rewards-math-train
Text Classification • 0.4B • Updated • 6
tuhink/hacking-rewards-code-train
Text Classification • 0.4B • Updated • 13
tuhink/hacking-rewards-helpful-train
Text Classification • 0.4B • Updated • 10
tuhink/hacking-rewards
Updated