Adnan KhanAdnanProduct & AI
San Francisco Bay Area
Get in touch

Source note2 min read

Neuron-Aware Active Few-Shot Learning for LLMs

Source

arXiv

arxiv.org/abs/2607.02423v1

Read the source

Source

Plain-Language Brief

The paper introduces a new method called NeuFS for Active Few-Shot Learning (AFSL) in Language Models (LLMs). Instead of relying on output-level signals to identify valuable unlabeled samples, it focuses on the internal dynamics of the model. The proposed method uses neuron activation patterns and a dual-criteria selection strategy to select diverse and informative samples, outperforming existing AFSL baselines in both reasoning and text classification tasks.

What the Document Says

  • Active Few-Shot Learning (AFSL) adapts LLMs to specialized domains by selecting valuable unlabeled samples for annotation.
  • Existing methods for sample identification rely on output-level signals like predictive entropy or semantic similarities based on external embeddings.
  • The proposed NeuFS framework shifts the selection paradigm to models' internal dynamics using neuron activation patterns.
  • NeuFS includes a dual-criteria selection strategy: ensuring diversity and broader example coverage, and prioritizing informative and challenging samples.
  • The strategy quantifies neuron consensus to identify samples where LLMs tend to hallucinate.
  • Experiments on three datasets show that NeuFS excels in both reasoning and text classification tasks, outperforming existing AFSL baselines.
  • Ablation studies highlight the effectiveness of internal neuron activations over external embeddings.

Key Concepts Mentioned

  • Active Few-Shot Learning
  • Language Models
  • Neuron-Aware Active Few-Shot Learning
  • neuron activation patterns
  • dual-criteria selection strategy
  • predictive entropy
  • semantic similarities
  • external embeddings

Claims Made by the Document

  • AFSL reduces human annotation costs.
  • NeuFS outperforms existing AFSL baselines in both reasoning and text classification tasks.
  • Internal neuron activations provide a more principled and effective selection signal than external embeddings.

Evidence Used

  • The text mentions 'reducing human annotation costs' as a benefit of AFSL.
  • Experiments on three datasets demonstrate NeuFS outperforming existing AFSL baselines.
  • Ablation studies support the effectiveness of internal neuron activations.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

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.