paper-with-me

홈 › Papers

Do Attention Heads Compete or Cooperate during Counting?

2025-02-10 · Pál Zsámboki, Ádám Fraknói, Máté Gedeon, András Kornai, Zsolt Zombori

We present an in-depth mechanistic interpretability analysis of training small transformers on an elementary task, counting, which is a crucial deductive step in many algorithms. In particular, we investigate the collaboration/competition among the attention heads: we ask whether the attention heads behave as a pseudo-ensemble, all solving the same subtask, or they perform different subtasks, meaning that they can only solve the original task in conjunction. Our work presents evidence that on the semantics of the counting task, attention heads behave as a pseudo-ensemble, but their outputs need to be aggregated in a non-uniform manner in order to create an encoding that conforms to the syntax. Our source code will be available upon publication.

📄 PDF Abstract BibTeX arXiv:2502.06923

Code (0)

등록된 구현이 없습니다.

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

SOFA-Net: Second-Order and First-order Attention Network for Crowd Counting

2020-08-09 · Haoran Duan, Shidong Wang, Yu Guan

Automated crowd counting from images/videos has attracted more attention in recent years because of its wide application in smart cities. But modelling the dense crowd heads is challenging and most of the existing works …

Crowd Counting

Effects of Parameter Norm Growth During Transformer Training: Inductive Bias from Gradient Descent

2020-10-19 · EMNLP 2021 11 · William Merrill, Vivek Ramanujan, Yoav Goldberg, Roy Schwartz 외

The capacity of neural networks like the widely adopted transformer is known to be very high. Evidence is emerging that they learn successfully due to inductive bias in the training routine, typically a variant of gradie…

Inductive Bias

An Adversarial Example for Direct Logit Attribution: Memory Management in GELU-4L

2023-10-11 · Jett Janiak, Can Rager, James Dao, Yeu-Tong Lau

Prior work suggests that language models manage the limited bandwidth of the residual stream through a "memory management" mechanism, where certain attention heads and MLP layers clear residual stream directions set by e…

Management

Agentic Transformers Provably Learn to Search via Reinforcement Learning

2026-05-29 · Tong Yang, Yu Huang, Yingbin Liang, Yuejie Chi arxiv

Tree search is a central abstraction behind many language-agent reasoning and decision-making tasks: agents must explore actions, remember failures, and backtrack toward promising alternatives. Yet, we lack a theoretical…

Reinforcement Learning

Attention to Head Locations for Crowd Counting

2018-06-27 · Youmei Zhang, Chunluan Zhou, Faliang Chang, Alex C. Kot

Occlusions, complex backgrounds, scale variations and non-uniform distributions present great challenges for crowd counting in practical applications. In this paper, we propose a novel method using an attention model to …

Crowd CountingDensity Estimation