·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Statistical Detection of Positional Bias in LLM-Based Recommender Systems: Methodologies and Empirical Evidence5h◆Self-Supervised Latent Action-Conditioned World Modeling for Time Series5h◆Merge and Branch: Efficient Tree-Structured Supervised Fine-Tuning of LLMs for Recommendation5h◆METATRON: A Promptable Whole-Body Execution Layer for Humanoids5h◆Estimating Implicit Expected Log Density Ratios5h◆Scaling Lean Theorem Proving for Real-World Code Verification5h◆A Tutorial on Generative Models for Music Audio5h◆Cross Lingual Abstractive Summarization for Marathi5h◆Persona Learning from Feedback for Personalized AI Assistants5h◆Learning Scientific Trajectories to Predict Future Biomedical Discoveries5h◆inkGround: Benchmarking Fine-grained Visual Grounding Capabilities of Vision-Language Models for Cursive Handwriting in the Wild5h◆Lev-3B: A Data Recipe for Compact Probabilistic Log Decision Models5h◆TableTune: Measured, Self-Correcting Draft-Length Tables for Dynamic Speculative Decoding5h◆Why Speculative Decoding Works Better Than Predicted on Sparse MoE Models5h◆Assessing feature scorer results on high-dimensional datasets with t-SNE5h◆Implicit Reasoning in Transformers: Bridging the Compositionality Gap in GPT-2-like Models5h◆Distilling the Shape of Uncertainty: What On-Policy Distillation Learns and How to Predict It5h◆Representation Complexity Along Learning Trajectories: Trading Width for Precision and Time5h◆PCM: A Proof-Carrying Message Protocol for Verifiable Multi-Agent Communication5h◆From Local Recovery to Unified Generation: Discovering Triple Set in Knowledge Graphs via Observation-Anchored Latent Diffusion5h◆Statistical Detection of Positional Bias in LLM-Based Recommender Systems: Methodologies and Empirical Evidence5h◆Self-Supervised Latent Action-Conditioned World Modeling for Time Series5h◆Merge and Branch: Efficient Tree-Structured Supervised Fine-Tuning of LLMs for Recommendation5h◆METATRON: A Promptable Whole-Body Execution Layer for Humanoids5h◆Estimating Implicit Expected Log Density Ratios5h◆Scaling Lean Theorem Proving for Real-World Code Verification5h◆A Tutorial on Generative Models for Music Audio5h◆Cross Lingual Abstractive Summarization for Marathi5h◆Persona Learning from Feedback for Personalized AI Assistants5h◆Learning Scientific Trajectories to Predict Future Biomedical Discoveries5h◆inkGround: Benchmarking Fine-grained Visual Grounding Capabilities of Vision-Language Models for Cursive Handwriting in the Wild5h◆Lev-3B: A Data Recipe for Compact Probabilistic Log Decision Models5h◆TableTune: Measured, Self-Correcting Draft-Length Tables for Dynamic Speculative Decoding5h◆Why Speculative Decoding Works Better Than Predicted on Sparse MoE Models5h◆Assessing feature scorer results on high-dimensional datasets with t-SNE5h◆Implicit Reasoning in Transformers: Bridging the Compositionality Gap in GPT-2-like Models5h◆Distilling the Shape of Uncertainty: What On-Policy Distillation Learns and How to Predict It5h◆Representation Complexity Along Learning Trajectories: Trading Width for Precision and Time5h◆PCM: A Proof-Carrying Message Protocol for Verifiable Multi-Agent Communication5h◆From Local Recovery to Unified Generation: Discovering Triple Set in Knowledge Graphs via Observation-Anchored Latent Diffusion5h◆
News/Distilling the Shape of Uncertainty: What On-Policy Distillation Learns and How to Predict It
arxiv
PublishedOctober 31, 2026 at 4:00 AM

Distilling the Shape of Uncertainty: What On-Policy Distillation Learns and How to Predict It

Source
arxiv.orgfull article ↗
Read on arxiv→
Publisher summary· verbatim

arXiv:2610.39020v1 Announce Type: cross Abstract: On-policy distillation (OPD) is increasingly adopted in the post-training of large language models. However, a stronger teacher may fail to improve the student. Prior work attributes this to a lack of new knowledge, but what that knowledge is remains

Stay posted· Newsletter

Three short emails a week: Monday, Wednesday and Friday. Top movers, new models and the most-referenced story.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Discussion
Source
↗
arxiv
Read original ↗All from arxiv →

No replies yet. Be first.

Source
↗
arxiv
Read original ↗All from arxiv →

Related coverage

More from ARXIV
arxivStatistical Detection of Positional Bias in LLM-Based Recommender Systems: Methodologies and Empirical Evidence5harxivSelf-Supervised Latent Action-Conditioned World Modeling for Time Series5harxivMerge and Branch: Efficient Tree-Structured Supervised Fine-Tuning of LLMs for Recommendation5harxivMETATRON: A Promptable Whole-Body Execution Layer for Humanoids5h
The Bubble Brief
MON · WED · FRI

Read AI insights every Monday, Wednesday and Friday — top movers, new models, story of the week.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Originally published on arxiv ↗
PrivacyTermsRefundsLegal notice|AboutMethodologyContact

© 2026 Databubble

Built by Marouane Gazouzi · Readyo, AI interview coach

HomeModelsNews