AI Paper Atlas

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

Reconstruction Attack-Resistant Inference Paradigm for LLM Cloud Services

Zipeng Ye, Wenjian Luo, Qi Zhou, Yubo Tang

AAAI 2026 · Large language models › Language, safety and interpretability · approach: LLM unlearning and benchmarks · 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
apply scaling defense to contextLlama-3-8B · HellaSwag (0-shot)accuracy: 66.7 -> 65.4Hurt
apply scaling defense to contextPhi-3-14B · GSM8K (0-shot CoT)accuracy: 91.1 -> 85.2Hurt
apply scaling defense to contextLlama-3-8B · HumanEval (pass@1)pass@1: 55.5 -> 51.2Hurt
add 8-bit quantization of user-side layers (ablation)defense, Mistral-7B · GSM8K (0-shot CoT)accuracy: 46.8 -> 46.6No clear effect
add 4-bit quantization of user-side layers (ablation)defense, Phi-3-14B · GSM8K (0-shot CoT)accuracy: 85.2 -> 78.1Hurt
add 4-bit quantization of user-side layers (ablation)defense, Mistral-7B · BoolQ (0-shot)accuracy: 82.8 -> 42.6Hurt
perturb embeddings instead of scaling defense (ablation)Mistral-7B · GSM8K (0-shot CoT)accuracy: 46.8 -> 2.1Hurt
replace tokens with nearest neighbors instead of scaling defense (ablation)Llama-3-8B · HumanEval (pass@1)pass@1: 51.2 -> 0Hurt
add defense to 3-shot examples vs 1-shot undefendedLlama-3-8B · BBH CoT (selected tasks)accuracy: 57.4 -> 67.5Improved
apply defense to MMLU few-shot examplesLlama-3-70B-AWQ · MMLUaccuracy (STEM): 77.7 -> 74.0Hurt

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