Papers
arxiv:2601.21996

Mechanistic Data Attribution: Tracing the Training Origins of Interpretable LLM Units

Published on Jan 29
ยท Submitted by
Jianhui Chen
on Jan 30
Authors:
,
,

Abstract

Mechnistic Data Attribution framework traces interpretable units to specific training samples using influence functions, demonstrating causal relationships between data structure and neural circuit formation in language models.

AI-generated summary

While Mechanistic Interpretability has identified interpretable circuits in LLMs, their causal origins in training data remain elusive. We introduce Mechanistic Data Attribution (MDA), a scalable framework that employs Influence Functions to trace interpretable units back to specific training samples. Through extensive experiments on the Pythia family, we causally validate that targeted intervention--removing or augmenting a small fraction of high-influence samples--significantly modulates the emergence of interpretable heads, whereas random interventions show no effect. Our analysis reveals that repetitive structural data (e.g., LaTeX, XML) acts as a mechanistic catalyst. Furthermore, we observe that interventions targeting induction head formation induce a concurrent change in the model's in-context learning (ICL) capability. This provides direct causal evidence for the long-standing hypothesis regarding the functional link between induction heads and ICL. Finally, we propose a mechanistic data augmentation pipeline that consistently accelerates circuit convergence across model scales, providing a principled methodology for steering the developmental trajectories of LLMs.

Community

We introduce Mechanistic Data Attribution (MDA), a new paradigm that shifts the focus of mechanistic interpretability from post-hoc circuit analysis to the causal formation of these mechanisms during training.

interesting

arXivLens breakdown of this paper ๐Ÿ‘‰ https://arxivlens.com/PaperView/Details/mechanistic-data-attribution-tracing-the-training-origins-of-interpretable-llm-units-54-66110475

  • Executive Summary
  • Detailed Breakdown
  • Practical Applications

This is an automated message from the Librarian Bot. I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API

Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on Hugging Face checkout this Space

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: @librarian-bot recommend

Sign up or log in to comment

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2601.21996 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2601.21996 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2601.21996 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.