AI Paper Atlas

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

Anchored Decoding: Provably Reducing Copyright Risk for Any Language Model

Jacqueline He, Jonathan Hayase, Wen-tau Yih, Sewoong Oh, Luke Zettlemoyer, Pang Wei Koh

ICML 2026 · Large language models › Language, safety and interpretability · approach: Parallel and early-exit decoding · 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
use Anchored Decoding instead of TokenSwap at the high-protection pointTinyComma 1.8B + Llama 3.1 70B · BIOS / BOOKSfactuality / fluency: 0.44 -> 0.53 factuality; 3.77 -> 4.02 fluencyImproved
use AnchoredByte Decoding instead of CP-Fuse for cross-tokenizer pairComma 7B + Llama 3.1 70B · BIOS / BOOKSfactuality / fluency: 0.23 -> 0.52 factuality; 3.75 -> 4.23 fluencyImproved
add system prompt telling the model to avoid copyrighted textRisky LM · BOOKSnormalized copying reductionNo clear effect
replace KL objective with infinity-Renyi divergence (ablation)Anchored Decoding · BOOKS / BIOS trade-offfluency vs factualityMixed
replace optimization with hard switch between risky and safe (NoOpt) (ablation)Anchored Decoding · BOOKS / BIOS trade-offrisk-utility ParetoHurt
replace adaptive start with safe-only cold start (ColdStart) (ablation)Anchored Decoding · BOOKS / BIOS trade-offrisk-utility ParetoHurt
aggregate prefix LLR by average instead of top-n (AvgDebt) (ablation)Anchored Decoding · BOOKS / BIOS trade-offrisk-utility ParetoHurt
remove prefix debt (NoDebt) (ablation)Anchored Decoding · BOOKS / BIOS trade-offrisk-utility ParetoHurt
use fixed per-step budget (Fixed) or lump-sum budget (Global) instead of adaptive banking (ablation)Anchored Decoding · BOOKS / BIOS trade-offrisk-utility ParetoHurt
pair the risky 70B model with a small 1.8B safe modelAnchored Decoding · Token-level wall-clockthroughput slowdown: 1.1x slower than Llama 3.1 70B alone vs RCAD 2.0xImproved
change debt window n (ablation)Anchored Decoding · BOOKSrisk-utilityNo clear effect

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