We have 7 of 23 papers
We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.
Not the right person? Other researchers publish under this name.
R-OPD, a gradient-triggered curriculum that adaptively selects when to revisit initial-student trajectories, improves average accuracy over continued current-policy sampling and produces longer responses, suggesting that well-timed revisits help students use more of their reasoning capacity.
RouteOPD (Routed On-Policy Distillation), which decomposes local teacher--student disagreement into student-excess sources and teacher-deficit destinations and couples them into explicit transport pairs and demonstrates the effectiveness of explicitly modeling probability transport in on-policy distillation.
It is shown the contamination is structurally irreversible: removing a source skill after the fact cannot erase the flawed reasoning its descendants have already inherited, so post-hoc rollback recovers only a small fraction of the lost performance, making skill admission a pre-commit necessity rather than a post-hoc f...
SPARK is introduced, which uses hidden-state response to diagnose whether a model internally enters an effective reasoning state and to guide lightweight test-time steering, and suggests that susceptibility can serve not only as a diagnostic signal for reasoning failures, but also as a practical guide for targeted test...
BIRD(Bootstrapped Iterative Self-Reasoning Distillation), a two-stage self-reasoning distillation method that improves the rollout distribution before on-policy training and achieves a stronger accuracy-efficiency trade-off than prompting and cold-start on-policy distillation on MATH-500 and AIME benchmarks.
This work proposes FADE (FFN Attenuation for DEcoding), a training-free method that attenuates FFN outputs to reduce language-prior dominance and shows that FADE effectively mitigates hallucinations while preserving inference efficiency.
SeeMe is proposed, a training-free framework that introduces the concept of feature engineering from traditional machine learning into LVLMs and restructures visual tokens through a three-stage token engineering process to suppress hallucination sources while preserving informative visual evidence.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
