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 research presents NeuFS, a new approach for Active Few-Shot Learning (AFSL) that focuses on using a model's internal dynamics to identify valuable unlabeled samples for annotation. This method is designed to reduce human annotation costs and improve performance by selecting samples based on neuron activation patterns rather than output-level proxies. The results show that NeuFS outperforms existing AFSL baselines in both reasoning and text classification tasks across three datasets.

What the Document Says

  • The research introduces a new method called Neuron-Aware Active Few-Shot Learning (NeuFS).
  • NeuFS aims to identify valuable unlabeled samples for annotation in specialized domains, reducing human annotation costs and promoting high performance.
  • Existing AFSL methods typically rely on output-level signals for sample identification.
  • However, these methods often overlook a model's internal dynamics that could pinpoint specific knowledge gaps.
  • NeuFS shifts the selection paradigm from output-level proxies to models' internal dynamics using neuron activation patterns.
  • It includes a dual-criteria selection strategy: ensuring few-shot sample diversity and prioritizing informative and challenging samples that LLMs tend to hallucinate.
  • NeuFS excels in both reasoning and text classification tasks, outperforming existing AFSL baselines across three datasets.
  • Ablation studies highlight that internal neuron activations provide a more principled and effective selection signal than external embeddings.

Key Concepts Mentioned

  • Active Few-Shot Learning (AFSL)
  • LLMs
  • neuron activation patterns
  • few-shot samples
  • hallucination

Claims Made by the Document

  • NeuFS excels in both reasoning and text classification tasks.
  • NeuFS outperforms existing AFSL baselines.

Evidence Used

  • Experiments on three datasets demonstrate that NeuFS excels in both reasoning and text classification tasks, outperforming existing AFSL baselines.

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.