Ya-Fu Li

Even top LLMs achieve full scientific closure in less than a third of physics research tasks, revealing a critical gap between generating evidence and drawing robust conclusions.

Value Flattening is identified as an important yet overlooked failure mode of critic learning in standard PPO and a simple sparse supervision strategy can mitigate it; SParse Proximal Policy Optimization is introduced, which applies the value loss to only a few well-separated states in each response to mitigate both effects.

Short-context models can achieve superior reasoning performance by leveraging long-context teacher models through innovative token alignment and training strategies.

GPT-5.5 not only tops the leaderboard in policy evolution but also reveals critical insights into how agents can optimize performance through strategic feedback utilization.

Even the best LLMs struggle with Olympiad-level combinatorics, achieving only 65.4% on a benchmark designed to expose their reasoning limitations.

Current personal assistant agents struggle to anticipate and act on unstated user needs in long, complex workflows, revealing a critical gap between task completion and genuine proactivity.

LVLMs can maintain sharper visual focus during long-form generation by adding a lightweight, learnable memory module that bypasses attention dilution.

Reasoning SFT doesn't just memorize, it generalizes—but only if you train it long enough, feed it good data, and use a capable model, and even then, reasoning gains come at the cost of safety.