AI Paper Atlas

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

Evaluating Text Creativity across Diverse Domains: a Dataset and Large Language Model Evaluator

Qian Cao, Xiting Wang, Yuzhuo Yuan, Yahui Liu, Fang Luo, Ruihua Song

ICLR 2026 · Large language models › Language, safety and interpretability · approach: LLM unlearning and benchmarks

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
train on CreataSet pairwise dataQwen2.5-7B-Instruct · CreataSet testagreement rate: 0.574 -> 0.745Improved
train on CreataSet pairwise dataQwen2.5-14B-Instruct · CreataSet testagreement rate: 0.661 -> 0.762Improved
train on CreataSet pairwise dataQwen2.5-14B-Instruct · CreataSet test, LyricsF1: -2.1%Hurt
provide a shared instruction as context for annotatorshuman annotation · 400 text pairsICC: 0.59 -> 0.75Improved
add swapped-position augmentation in training (ablation)CrEval-7B · CreataSet testconsistencyImproved
add human original responses to training data (ablation)CrEval · CreataSet testagreementImproved
add synthetic responses to training data (ablation)CrEval · CreataSet testagreementImproved
add negative sampling of unrelated responses (ablation)CrEval · CreataSet testagreementImproved
increase training data beyond 100K samples (ablation)CrEval · CreataSet testF1/Kappa/agreementNo clear effect
use classification loss instead of Bradley-Terry loss (ablation)CrEval-7B · CreataSet testconsistency: 0.846 -> 0.920Improved
use Qwen2.5-14B instead of 7B backbone (ablation)CrEval · CreataSet testagreement rate: 0.745 -> 0.762Improved
replace contextual DPO rejects with random negative rejects (ablation)DPO on Qwen2.5-7B-Instruct · win rate vs GPT-4o-miniwin rateHurt

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