AI Paper Atlas

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

Learning-Time Encoding Shapes Unlearning in LLMs

Ruihan Wu, Konstantin Garov, Kamalika Chaudhuri

ICLR 2026 · Large language models › Language, safety and interpretability · approach: LLM unlearning and benchmarks · 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
fine-tune with paraphrased descriptions (FT-Mul) instead of single descriptionsunlearning with GA and TV · Eval-DU+ and TOFU+, extraction trade-offNorm-AUC / AUC: FT-Mul outperforms or matches FT-Single in 44 out of 48 casesImproved
fine-tune with paraphrased descriptions (FT-Mul)TV UL-Exact, Llama2-7B · Eval-DU+ extractionNorm-AUC: 0.57 -> 0.65Improved
fine-tune with paraphrased descriptions (FT-Mul)GA UL-Exact, Llama2-7B · Eval-DU+ extractionNorm-AUC: 0.59 -> 0.55Hurt
fine-tune with paraphrased descriptions (FT-Mul)TV UL-Mul, Llama2-7B · Eval-DU+ extractionNorm-AUC: 0.69 -> 0.72Improved
fine-tune with paraphrased descriptions (FT-Mul)GA UL-Exact, Llama3-8B · Eval-DU+ extractionNorm-AUC: 0.55 -> 0.62Improved
fine-tune with paraphrased descriptions (FT-Mul)GA UL-Exact, Gemma2-2B · Eval-DU+ extractionNorm-AUC: 0.52 -> 0.60Improved
fine-tune with paraphrased descriptions (FT-Mul)FT-Mul versus FT-Single, memorization trade-off · Eval-DU+ and TOFU+, memorization trade-offNorm-AUC / AUC: outperforms or matches FT-Single in 39 out of 48 casesImproved
align the unlearn split with chunk boundaries of the training textFT-Mul-Chunk, GA UL-Exact, Llama2-7B · Eval-DU+ extractionNorm-AUC: 0.53 -> 0.66 (+0.13)Improved
align the unlearn split with chunk boundariesFT-Mul-Chunk, TV UL-Exact, Gemma2-2B · Eval-DU+ extractionNorm-AUC: 0.61 -> 0.63 (+0.02)Improved
isolate target knowledge by concatenating individual descriptions into chunks (FT-Mul-Chunk-Iso)GA UL-Single, Llama2-7B · Eval-DU+ extractionNorm-AUC: +0.11Improved
isolate target knowledge in chunks (FT-Mul-Chunk-Iso)TV UL-Mul, Llama2-7B · Eval-DU+ extractionNorm-AUC: +0.13Improved
organize unlearn targets by author partitions (TOFU+) rather than entangled chunks (Eval-DU+)GA UL-Exact, Llama2-7B · extraction trade-offNorm-AUC: 0.53 -> 0.59Improved

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