Interesting Work

The three works below are interesting to me for the same reason: they use toy models and careful experimental design to make reasoning behavior more interpretable.

• RL Grokking Recipe: How Does RL Unlock and Transfer New Algorithms in LLMs?
• On the Interplay of Pre-Training, Mid-Training, and RL on Reasoning Language Models
• Is Chain-of-Thought Reasoning of LLMs a Mirage? A Data Distribution Lens

Friendship

• Haoyu Zhang A friend with a calm sense of direction who always gives people the feeling that things are moving forward in a steady way.

• Yulong Chen A sincere and hardworking friend who is full of energy, easy to talk to, and always brings a grounded presence.

• Changkang Li A friend and infra engineer who brings real passion and insight to everyday life.

Life

I also want this site to keep a light personal corner beyond papers and code. Over time, I plan to use this page for a few daily-life photos, short notes, and small snapshots of the moments around research life.

For now, this page serves as a quiet placeholder for that side of the site while the academic sections continue to grow.