Neuron-Aware Active Few-Shot Learning for LLMs
2026-07-02 • Machine Learning
Machine LearningArtificial Intelligence
AI summaryⓘ
The authors propose NeuFS, a method to improve how large language models (LLMs) learn from a small number of examples. Instead of picking examples based on the model's final answers or external measures, NeuFS looks inside the model at how its neurons activate. This helps find diverse and challenging examples that the model struggles with, making learning more effective. Tests on different tasks show NeuFS works better than previous methods that use output-level signals.
Active Few-Shot LearningLarge Language ModelsNeuron ActivationSample SelectionModel Internal DynamicsPredictive EntropySemantic SimilarityHallucinationReasoning TasksText Classification
Authors
Zhuowei Chen, Liwei Chen, Christian Schunn, Raquel Coelho, Xiang Lorraine Li
Abstract
Active Few-Shot Learning (AFSL) adapts LLMs to specialized domains by identifying the most valuable unlabeled samples for annotation and use as few-shot demonstrations, effectively reducing human annotation costs while promoting high performance. However, existing methods typically rely on output-level signals for sample identification, such as predictive entropy or semantic similarities with test-time data based on external embeddings, which often overlook models' internal dynamics, which could pinpoint specific knowledge gaps. To bridge this gap, we propose NeuFS, a Neuron-Aware Active Few-Shot Learning framework that shifts the selection paradigm from output-level proxies to models' internal dynamics. NeuFS utilizes neuron activation patterns to represent sample directly, and includes a dual-criteria selection strategy that: (1) ensures few-shot sample diversity with neuron patterns for broader example coverage, while (2) prioritizing on identifying informative and challenging few-shot samples LLMs tend to hallucinate by quantifying neuron consensus. Experiments on three datasets demonstrate that NeuFS excels in both reasoning and text classification tasks, outperforming existing AFSL baselines. Ablation studies further highlight that internal neuron activations provide a more principled and effective selection signal than external embeddings, validating the superiority of the proposed NeuFS.