Self-Policy Distillation via Capability-Selective Subspace Projection

arXiv:2605.22675 · cs.CL · Submitted 2026-05-21 · Read on arXiv

cs.CL

Submitted: 2026-05-21

Updated: 2026-10-11

Code: https://github.com/sahil280114/codealpaca

Terminology

• Distillation — 4× in this paper · explained in 2 episodes
• Self-Distillation — 2× in this paper · explained in 7 episodes
• Self-Policy Distillation — 2× in this paper
• Generalizability — 1× in this paper · explained in 6 episodes
• Generalization — 1× in this paper · explained in 13 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes

Sources

• Beyond Human Data: Scaling Self-Training for Problem-Solving with Language Models
• Embarrassingly Simple Self-Distillation Improves Code Generation
• Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe
• Style over Substance: Distilled Language Models Reason Via Stylistic Replication
• V-STaR: Training Verifiers for Self-Taught Reasoners
• Privileged Information Distillation for Language Models
• A Survey of On-Policy Distillation for Large Language Models
• Retaining by Doing: The Role of On-Policy Data in Mitigating Forgetting
• Towards Cross-Tokenizer Distillation: the Universal Logit Distillation Loss for LLMs
• A General Language Assistant as a Laboratory for Alignment
• RLEF: Grounding Code LLMs in Execution Feedback with Reinforcement Learning
• Training Large Language Models to Reason in a Continuous Latent Space
• KV-Distill: Nearly Lossless Learnable Context Compression for LLMs
• System-1.5 Reasoning: Traversal in Language and Latent Spaces with Dynamic Shortcuts
• Program Synthesis with Large Language Models
• Measuring Massive Multitask Language Understanding
• Qwen2.5 Technical Report
• Qwen3 Technical Report

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering