arXiv AI
Distilling Directional Verification
Read the original on arXiv AI →The Flow has not summarised this story yet — read it at arXiv AI.
Related stories
arXiv Computation and Language
Sep 2
Knowledge Distillation During Mid-Training Favors Reasoning over Factual Recall
arXiv:2609.01532v1 Announce Type: new Abstract: Logit-based knowledge distillation (KD) is used to train smaller language models (LMs) via supervision from stronger teachers, but whether its benefits...
arXiv Machine Learning
1d ago
Activation-Conditioned Self-Distillation
arXiv:2609.38342v1 Announce Type: new Abstract: On-policy self-distillation uses a model as its own teacher to provide dense supervision for reasoning, often through reference-solution conditioning....
arXiv Computation and Language
2d ago
Training LLM Judges from Language Feedback via Position-Selective Self-Distillation
arXiv:2609.38792v1 Announce Type: new Abstract: We study training LLM judges from natural language feedback, especially for subjective tasks where the verdict depends strongly on which evaluation cri...
arXiv Machine Learning
Jul 9
Geometric Self-Distillation for Reasoning Generalization
arXiv:2607. 06855v1 Announce Type: new Abstract: On-policy distillation is a practical post-training recipe for large language models, supplying dense teacher supervision on the student's own trajectories.
arXiv Computation and Language
3d ago
Distilling What Matters: Confidence-Aware Selective Distillation for Large Language Models
arXiv:2609.36734v1 Announce Type: new Abstract: Knowledge Distillation (KD) trains a smaller-capacity student model to imitate a larger-capacity teacher model by matching output distributions, implic...
arXiv Machine Learning
Aug 20
Rethinking Privileged Information in On-Policy Self-Distillation
The paper investigates on‑policy self‑distillation (OPSD), where a student model learns from its own outputs using token‑level supervision conditioned on privileged reference information. Experiments with Qwen3 models on science and mathematics datasets show that the correct reference does not consistently improve performance; students can improve without it, and solutions from other problems sometimes outperform the correct reference. The study finds that student predictions align more closely with the base model’s reasoning than with the reference supervision, and that alignment alone does not reliably predict performance gains.