AI Paper Atlas

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

Rethinking LLM-as-a-Judge: Representation-as-a-Judge with Small Language Models via Semantic Capacity Asymmetry

Zhuochun Li, Yong Zhang, Ming Li, Yuelyu Ji, Yiming Zeng, Ning Cheng, Yun Zhu, Yanmeng Wang, Shaojun Wang, Jing Xiao, Daqing He

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Pruning and circuits in LLMs · 5 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
probe hidden states instead of prompting the small LMQwen3-1.7B · MATH Logicality (binary)weighted F1 (%): 51.34 -> 83.67Improved
probe hidden states instead of prompting the small LMQwen3-1.7B · MATH Logicality (multiclass 1-5)weighted F1 (%): 15.06 -> 59.05Improved
probe hidden states instead of prompting the small LMLlama-3.1-8B-Instruct · MATH Informativeness (binary)weighted F1 (%): 86.67 -> 88.12Improved
use mean pooling instead of other pooling strategies (ablation)INSPECTOR probe · MATH Informativeness (binary)F1Improved
use logistic regression instead of other classifiers (ablation)INSPECTOR probe · MATH Informativeness (binary)F1Improved
use a larger small LM (Llama-3.1-8B vs Qwen3-1.7B) (ablation)INSPECTOR probe · reasoning benchmarksweighted F1 (%): 96.32 vs 92.65 (Llama-3.2-1B vs Llama-3.1-8B, MATH Fluency binary)Mixed
train probe on one dataset and test on another (OOD) vs in-distribution (ablation)INSPECTOR probe Qwen3-1.7B · GSM8K/MATH multiclassweighted F1 (%): multiclass F1 drops to roughly 10-25%Hurt
filter SFT data with probing classifier instead of randomLlama-2-7B-Chat distillation · reasoning benchmarks SFTSFT performanceImproved

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