Distillation Defenses Easily Break After Reinforcement Learning
cs.LG, cs.AI, cs.CR
Submitted: 2026-09-28
Updated: 2026-09-28
Code: https://github.com/sjavaheri/treason
Terminology
Sources
- Llama-Nemotron: Efficient Reasoning Models
- Evaluating Large Language Models Trained on Code
- Training Verifiers to Solve Math Word Problems
- Boundary Point Jailbreaking of Black-Box LLMs
- Gemma: Open Models Based on Gemini Research and Technology
- The Llama 3 Herd of Models
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- Distilling the Knowledge in a Neural Network
- Why Distillation can Outperform Zero-RL: The Role of Flexible Reasoning
- Reinforcement Learning and Control as Probabilistic Inference: Tutorial and Review
- Understanding R1-Zero-Like Training: A Critical Perspective
- AceReason-Nemotron 1.1: Advancing Math and Code Reasoning through SFT and RL Synergy
- Stealing Reasoning Traces from Proprietary LLM APIs
- Qwen2.5 Technical Report
- Proximal Policy Optimization Algorithms
- Spurious Rewards: Rethinking Training Signals in RLVR
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Qwen3 Technical Report
- SimpleRL-Zoo: Investigating and Taming Zero Reinforcement Learning for Open Base Models in the Wild
- How to Steal Reasoning Without Reasoning Traces
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks