AI Paper Atlas

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

Next Generation Active Learning: Mixture of LLMs in the Loop

Yuanyuan Qi, Xiaohao Yang, Jueqing Lu, Guoxiang Guo, Joanne Enticott, Gang Liu, Lan Du

AAAI 2026 · Large language models › Language, safety and interpretability · approach: Statistical frameworks for LLM evaluation · 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
add pseudo-labeling to MoLAM (ablation)MoLAM annotator · TRECannotation accuracy: 0.7924 -> 0.8040Improved
add pseudo-labeling to MoLAM (ablation)MoLAM annotator · AG Newsannotation accuracy: 0.8819 -> 0.8887Improved
add pseudo-labeling to MoLAM (ablation)MoLAM annotator · IMDBannotation accuracy: 0.9534 -> 0.9538No clear effect
use trained MoLAM instead of vote-based ensembleLLM annotation · TRECannotation accuracy: 0.7320 -> 0.8040Improved
use trained MoLAM instead of best single LLM (Yi)LLM annotation · AG Newsannotation accuracy: 0.8349 -> 0.8887Improved
use MoLAM instead of vocabulary-level data augmentation DAannotator · TRECannotation accuracy: 0.4186 -> 0.8040Improved
add negative learning to active learning (ablation)MoLLIA · active learning micro-F1micro-F1Improved
add annotation discrepancy weighting (ablation)MoLLIA · active learning micro-F1micro-F1Improved
use negative labels from MoLAM (ablation)MoLLIA · AG Newsfalse negative label proportion: 0.0030No 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