paper-with-me

홈 › Papers

Induction Heads Interpolate N-Grams

2026-07-02 · Francesco D'Angelo, Oguz Kaan Yuksel, Swathi Shree Narashiman, Nicolas Flammarion arxiv

Induction heads are attention circuits believed to underlie in-context learning in transformers, yet a precise characterization of the estimators they implement remains elusive. We study transformers trained on order-$k$ Markov chains and identify two complementary smoothing mechanisms. First, at finite attention-weight scale, the circuit implements a soft context-matching estimator: it aggregates contributions from exact and partial context matches, weighted exponentially by their overlap, and induces a data-dependent interpolation across context orders analogous to Jelinek-Mercer smoothing. Second, a beginning-of-sequence (BOS) token induces additive pseudo-counts, recovering Dirichlet-style smoothing. We construct a disentangled transformer implementing both mechanisms and show that trained transformers recover the predicted attention patterns. Across settings where pseudo-count smoothing is optimal or lower-order contexts provide structured evidence, trained transformers match or outperform classical count-based baselines. Our results bridge mechanistic interpretability of induction heads with classical statistical smoothing, revealing that transformers learn to regularize in-context estimation rather than simply count.

📄 PDF Abstract BibTeX arXiv:2607.02800

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

The Dual-Route Model of Induction

2025-04-03 · Sheridan Feucht, Eric Todd, Byron Wallace, David Bau

Prior work on in-context copying has shown the existence of induction heads, which attend to and promote individual tokens during copying. In this work we introduce a new type of induction head: concept-level induction h…

In-Context Learningmodel

In-Context Learning Without Copying

2025-11-07 · Kerem Sahin, Sheridan Feucht, Adam Belfki, Jannik Brinkmann 외 arxiv

Induction heads are attention heads that perform inductive copying by matching patterns from earlier context and copying their continuations verbatim. As models develop induction heads, they experience a sharp drop in tr…

Which Attention Heads Matter for In-Context Learning?

2025-02-19 · Kayo Yin, Jacob Steinhardt

Large language models (LLMs) exhibit impressive in-context learning (ICL) capability, enabling them to perform new tasks using only a few demonstrations in the prompt. Two different mechanisms have been proposed to expla…

In-Context Learning

In-context Learning and Induction Heads

2022-09-24 · Catherine Olsson, Nelson Elhage, Neel Nanda, Nicholas Joseph 외

"Induction heads" are attention heads that implement a simple algorithm to complete token sequences like [A][B] ... [A] -> [B]. In this work, we present preliminary and indirect evidence for a hypothesis that induction h…

In-Context Learning

The Evolution of Statistical Induction Heads: In-Context Learning Markov Chains

2024-02-16 · Benjamin L. Edelman, Ezra Edelman, Surbhi Goel, Eran Malach 외

Large language models have the ability to generate text that mimics patterns in their inputs. We introduce a simple Markov Chain sequence modeling task in order to study how this in-context learning (ICL) capability emer…

In-Context Learning