Yuntai Bao

colored-dye

·

https://colored-dye.github.io/

colored-dye

AI & ML interests

on-policy distillation, reinforcement learning, mechanistic interpretability, training data attribution

Recent Activity

upvoted a paper 4 days ago

DTCA: Denoising Token Credit Assignment for Reinforcement Learning in Diffusion Language Models

upvoted a collection 23 days ago

liked a model 23 days ago

Qwen/Qwen3.8-27B

View all activity

Organizations

None yet

Papers 5

arxiv:2609.32722

arxiv:2608.06880

arxiv:2607.21106

arxiv:2605.05974

models 2

colored-dye/OPD-scaling-checkpoints

Reinforcement Learning • Updated Sep 30 • 1

colored-dye/axbench-steering-vector

datasets 1

colored-dye/concept500-contrastive

Viewer • Updated May 10 • 144k • 45