AI Paper Atlas

Research map / Large language models / Instruction and fine-tuning

Inverse IFEval: Can LLMs Unlearn Stubborn Training Conventions to Follow Real Instructions?

Qinyan Zhang, Xinping Lei, Ruijie Miao, FU YU, Haojie Fan, Le Chang, Jiafan Hou, Dingling Zhang, Zhongfei Hou, ZiqiangYang, puchangxin, FEI HU and 8 more

ICLR 2026 · Large language models › Instruction and fine-tuning · approach: Efficient LLM fine-tuning systems

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 thinking mode instead of non-thinking mode (ablation)Qwen3 series · Inverse IFEval (English)overall score: Qwen3-235B-A22B 40.28 (Instruct) -> 54.22 (Thinking)Improved
use thinking mode instead of non-thinking mode (ablation)Qwen3-30B-A3B · Inverse IFEval (English)overall score: 30.43 -> 49.21Improved
thinking mode vs auto vs non-thinking (ablation)Doubao-Seed-1.6 (AdaCoT) · Inverse IFEvaloverall scoreImproved
use auto-thinking mode instead of non-thinking (ablation)Doubao-Seed-1.6 (AdaCoT) · Inverse IFEval (Chinese)overall scoreHurt
use full thinking budget (Pro) instead of reduced budget (Flash) (ablation)Gemini-2.5 · Inverse IFEval (English)overall score: 68.61 -> 70.55Improved
scale up model parameters (ablation)Qwen3 series · Inverse IFEvaloverall score: English Instruct 30.43 -> 40.28 (30B-A3B -> 235B-A22B)Improved
increase Best-of-N from 1 to 16 and 32 (ablation)five models · Inverse IFEval (avg EN+ZH)overall scoreImproved
evaluate counterintuitive instructions instead of conventional IFEval instructionsQwen3-235B-A22B-Instruct · IFEval vs Inverse IFEval rankingrank among 15 models: 5th -> 15thHurt
compare Chinese-developed models in Chinese vs EnglishGLM-4.5 and Qwen3-235B-A22B-Thinking · Inverse IFEvaloverall score: GLM-4.5 58.30 -> 66.96; Qwen3-235B-Thinking 54.22 -> 70.62Improved
optimize judge model, template and system prompt per instruction typeLLM-as-a-Judge · judge validationjudging accuracy: 88% -> 98%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