AI Paper Atlas

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

Diverse Text Decoding via Iterative Reweighting

Ruiqi Shi, Sinno Jialin Pan

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Parallel and early-exit decoding

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 OverRIDE iterative reweighting to greedy decodingQwen-2.5-7B-Instruct · HumanEvalPASS@10: 64.9 -> 86.0Improved
add OverRIDE to greedy decodingMistral-7B-Instruct-v0.3 · GSM8KPASS@10: 39.5 -> 74.2Improved
add OverRIDE to top-p sampling (tau=1.0)Qwen-2.5-7B-Instruct · MATH500PASS@10: 87.2 -> 87.6Improved
add OverRIDE to greedy decodingQwen-2.5-7B-Instruct · HumanEvalCodeBLEU (pairwise): 1.000 -> 0.626Improved
add OverRIDE to greedy decodingQwen-2.5-7B-Instruct · CCNewsMAUVE: 0.803 -> 0.753Hurt
add OverRIDE to top-p sampling (tau=0.6)Qwen-2.5-7B-Instruct · CCNewsMAUVE: 0.901 -> 0.938Improved
add OverRIDE to larger model (14B vs 3B)Qwen-2.5 · HumanEval PASS@10 (top-p tau=0.6)PASS@10: 3B: 82.9 -> 83.2; 14B: 71.1 -> 77.2Improved
use low-rank output head adapter rank 16 instead of full output heads (ablation)OverRIDE on Qwen-2.5-7B · MATH500throughput: 1984.1 -> 3435.8Improved
raise adapter rank from 16 to 256 (ablation)OverRIDE on Qwen-2.5-7B · MATH500PASS@10: 87.5 -> 87.1No clear effect
use rank-4 adapter instead of rank 16 (ablation)OverRIDE on Qwen-2.5-7B · MATH500PASS@10: 86.9 -> 87.5Improved
OverRIDE instead of Contrastive DecodingGPT2-XL · CCNewsMAUVE: 0.752 -> 0.846Improved
OverRIDE instead of SimCTGQwen-2.5-7B-Instruct · HumanEvalPASS@10: 77.2% -> 86.4%Improved

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