AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Uncovering Pretraining Code in LLMs: A Syntax-Aware Attribution Approach

Yuanheng Li, Zhuoyang Chen, Xiaoyun Liu, Yuhao Wang, Mingwei Liu, Yang Shi, Kaifeng Huang, Shengjie Zhao

AAAI 2026 · Large language models › Language, safety and interpretability · approach: Data and hallucination detection · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
prune syntax-determined tokens when scoring membership (SynPrune vs best baseline DC-PDD)MIA on Pythia 2.8B · Python function MIA benchmark (1:1)AUROC (%): 50.1 -> 61.3Improved
prune syntax-determined tokens (SynPrune vs LOSS)MIA on GPT-J 6B · Python function MIA benchmark (1:1)AUROC (%): 43.3 -> 59.7Improved
prune syntax-determined tokens at 5:1 member-heavy ratioMIA on GPT-J 6B · Python function MIA benchmark (5:1)AUROC (%): 43.6 -> 63.1Improved
add compound-statement syntax conventions to pruning (ablation)SynPrune on StableLM-Alpha 3B · Python function MIA benchmark (1:1)AUROC (points): -7.1 when removedImproved
add data-model syntax conventions to pruning (ablation)SynPrune on Pythia 2.8B · Python function MIA benchmark (1:1)AUROC (points): -5.1 when removedImproved
add expression syntax conventions to pruning (ablation)SynPrune on Pythia 2.8B · Python function MIA benchmark (1:1)AUROC (points): -4.5 when removedImproved
add single-statement syntax conventions to pruning (ablation)SynPrune on GPT-J 6B · Python function MIA benchmark (1:1)AUROC (points): -1.6 when removedImproved
add data-model conventions to pruning in GPT-Neo at 1:5 (ablation)SynPrune on GPT-Neo 2.7B · Python function MIA benchmark (1:5)AUROC (points): -0.1 when removedNo clear effect
evaluate on long instead of short member functionsSynPrune on GPT-J 6B · Python function MIA benchmark (member functions by length)false negative rate (%): 0.00 -> 87.09Hurt
evaluate on long instead of short member functionsSynPrune on Pythia 2.8B · Python function MIA benchmark (member functions by length)false negative rate (%): 2.10 -> 75.12Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy