paper-with-me

Papers

AttentionInfluence: Adopting Attention Head Influence for Weak-to-Strong Pretraining Data Selection

2025-05-12 · Kai Hua, Steven Wu, Ge Zhang, Ke Shen

Recently, there has been growing interest in collecting reasoning-intensive pretraining data to improve LLMs' complex reasoning ability. Prior approaches typically rely on supervised classifiers to identify such data, which requires labeling by humans or LLMs, often introducing domain-specific biases. Due to the attention heads being crucial to in-context reasoning, we propose AttentionInfluence, a simple yet effective, training-free method without supervision signal. Our approach enables a small pretrained language model to act as a strong data selector through a simple attention head masking operation. Specifically, we identify retrieval heads and compute the loss difference when masking these heads. We apply AttentionInfluence to a 1.3B-parameter dense model to conduct data selection on the SmolLM corpus of 241B tokens, and mix the SmolLM corpus with the selected subset comprising 73B tokens to pretrain a 7B-parameter dense model using 1T training tokens and WSD learning rate scheduling. Our experimental results demonstrate substantial improvements, ranging from 1.4pp to 3.5pp, across several knowledge-intensive and reasoning-heavy benchmarks (i.e., MMLU, MMLU-Pro, AGIEval-en, GSM8K, and HumanEval). This demonstrates an effective weak-to-strong scaling property, with small models improving the final performance of larger models-offering a promising and scalable path for reasoning-centric data selection.

📄 PDF Abstract BibTeX arXiv:2505.07293

Code (0)

등록된 구현이 없습니다.

Tasks

GSM8KHumanEvalMMLUScheduling

Methods 이 논문이 사용한 방법론

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Attention 설명 없음

Similar Papers 제목 키워드 기반

Dissecting Persona-Driven Reasoning in Language Models via Activation Patching

2025-07-28 · Ansh Poonia, Maeghal Jain arxiv

Large language models (LLMs) exhibit remarkable versatility in adopting diverse personas. In this study, we examine how assigning a persona influences a model's reasoning on an objective task. Using activation patching, …

AIR: Post-training Data Selection for Reasoning via Attention Head Influence

2025-12-15 · Jinrui Liu, Jeff Wu, Xuanguang Pan, Gavin Cheung 외 arxiv

LLMs achieve remarkable multi-step reasoning capabilities, yet effectively transferring these skills via post-training distillation remains challenging. Existing data selection methods, ranging from manual curation to he…

The Anxiety of Influence: Bloom Filters in Transformer Attention Heads

2026-02-19 · Peter Balogh arxiv

Some transformer attention heads appear to function as membership testers, dedicating themselves to answering the question "has this token appeared before in the context?" We identify these heads across four language mod…

SFC: Shared Feature Calibration in Weakly Supervised Semantic Segmentation

2024-01-22 · Xinqiao Zhao, Feilong Tang, Xiaoyang Wang, Jimin Xiao

Image-level weakly supervised semantic segmentation has received increasing attention due to its low annotation cost. Existing methods mainly rely on Class Activation Mapping (CAM) to obtain pseudo-labels for training se…

Pseudo LabelSegmentationSemantic SegmentationWeakly supervised Semantic Segmentation+1

Influence Score and Transformers interpretability: Measure of the Effective Impact of Attention Heads at inference time

2026-09-04 · Lisa Bouger, Yannick Teglia, Philippe Loubet Moundi arxiv

We propose an influence score to quantify the contribution of attention heads to classification decisions in Transformer-based models designed for prompt injection detection. The score combines directional influence on t…