AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Provable Training Data Identification for Large Language Models

Zhenlong Liu, Hao Zeng, Weiran Huang, Hongxin Wei

ICML 2026 · Learning theory and algorithms › Prediction and decision losses · approach: Active learning and subset selection · 0 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
add p-value scaling with JKBB estimate of member proportion (ablation)conformal BH · WikiMIA NeoX-20B MIN-K% alpha=0.1power (%): 1.69 -> 5.90Improved
add p-value scaling with JKBB estimate of member proportion (ablation)conformal BH · WikiMIA NeoX-20B perplexity alpha=0.2power (%): 2.37 -> 10.88Improved
add p-value scaling with JKBB estimate of member proportion (ablation)conformal BH · WikiMIA NeoX-20B MIN-K% target FIR 0.5power: 0.5177 -> 0.8618Improved
use JKBB estimator instead of Storey estimator (ablation)PTDI · WikiMIA NeoX-20B MIN-K% alpha=0.3power (%): 42.81 -> 53.31Improved
use JKBB estimator instead of BKY (ablation)PTDI · WikiMIA LLaMA-7B MIN-K% alpha=0.3power (%): 8.47 -> 18.47Improved
use JKBB estimator instead of Quantile-BH (ablation)PTDI · WikiMIA Pythia-6.9B MIN-K% alpha=0.3power (%): 10.25 -> 24.24Improved
use PTDI instead of knockoff-based KTDFIR control · WikiMIA and XSum, white-box knockoff statisticFIRImproved
use PTDI instead of prior approachFIR control · WikiMIA Pythia-1.4B target FIR 0.05empirical FIR: 0.1311 -> 0.0494Improved
increase calibration-to-test ratio rho from 0.1 to 1 (ablation)PTDI · ArXivTectionFIP varianceImproved
vary training member proportion pi_test (ablation)PTDI · VL-MIA/Flickr MiniGPT-4FIRNo clear effect
add adjusted moment estimator using known members (ablation)PTDI · WikiMIApowerImproved

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