Post-Training Leaves Behavioral Shadows on Unrelated Decisions

arXiv:2609.29233 · cs.CL, cs.AI, cs.LG · Submitted 2026-09-24 · Read on arXiv

cs.CL, cs.AI, cs.LG

Submitted: 2026-09-24

Updated: 2026-09-24

Code: https://github.com/myboker/ATD

Terminology

• Distillation — 22× in this paper · explained in 2 episodes
• Alignment — 15× in this paper · explained in 1 episode
• Benchmark — 14× in this paper · explained in 2 episodes
• Adaptation — 10× in this paper · explained in 1 episode
• Active Learning — 4× in this paper · explained in 6 episodes
• Active Taskless Distillation — 4× in this paper
• Fine-Tuning — 4× in this paper · explained in 5 episodes
• On-Policy Distillation — 4× in this paper · explained in 3 episodes
• Robustness — 4× in this paper · explained in 5 episodes
• Knowledge Distillation — 3× in this paper · explained in 11 episodes
• Accuracy — 2× in this paper · explained in 1 episode
• Generalization — 2× in this paper · explained in 13 episodes
• Ablation — 1× in this paper · explained in 1 episode
• Core — 1× in this paper · explained in 2 episodes
• Elias — 1× in this paper · explained in 2 episodes
• Fingerprinting — 1× in this paper · explained in 2 episodes
• Instruction Tuning — 1× in this paper · explained in 2 episodes
• LoRA Adapters — 1× in this paper · explained in 3 episodes
• Steering — 1× in this paper · explained in 2 episodes
• Trace — 1× in this paper · explained in 2 episodes
• Verifier — 1× in this paper · explained in 2 episodes

Sources

• Subliminal Learning Is Steering Vector Distillation
• Subliminal Learning is Non-Semantic Distillation
• You Didn't Have to Say It like That: Subliminal Learning from Faithful Paraphrases
• Subliminal Signals in Preference Labels
• Scaling Model-Generated Distillation Data Can Make Latent Teacher Traits More Recoverable
• Distilling the Knowledge in a Neural Network
• Magicoder: Empowering Code Generation with OSS-Instruct
• Black-Box On-Policy Distillation of Large Language Models
• Subliminal Effects in Your Data: A General Mechanism via Log-Linearity
• When Do PEFT Adaptations Leak Structure? Measuring Black-Box Structural Bounds in Public-Base Model Services
• Editing Models with Task Arithmetic
• Qwen2.5 Technical Report
• LoRA: Low-Rank Adaptation of Large Language Models
• Evaluating Large Language Models Trained on Code
• Training Verifiers to Solve Math Word Problems
• 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