Skip to content
AI News HubLIVE
Original source2 min read

Pruning via Causal Attribution Preserves Reasoning Performance in Large Language Models

Summary

Researchers propose Causal Attribution Pruning (CAP), a training-free method that identifies critical attention heads in LLMs by measuring their causal impact on reasoning tasks. CAP achieves up to 61% relative accuracy improvement over Wanda at 20% sparsity on ARC-Challenge, and consistently outperforms at moderate sparsity levels.

SourcearXiv Computational LinguisticsAuthor: Amogh Sheth, Biruk Assefa, Yi Wen Huang, Andrew Lin, Yuhao Ge
Pruning via Causal Attribution Preserves Reasoning Performance in Large Language Models
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[2606.19350] Pruning via Causal Attribution Preserves Reasoning Performance in Large Language Models

[Submitted on 27 Apr 2026]

Title:Pruning via Causal Attribution Preserves Reasoning Performance in Large Language Models

View a PDF of the paper titled Pruning via Causal Attribution Preserves Reasoning Performance in Large Language Models, by Amogh Sheth and 4 other authors

View PDF HTML (experimental)

Abstract:Large language models (LLMs) excel at multi-step reasoning but incur substantial inference cost. We introduce Causal Attribution Pruning (CAP), a training-free method that identifies critical attention heads by measuring their causal impact on reasoning tasks and uses these head-level scores to guide fine-grained weight pruning. For each attention head, CAP estimates the expected performance degradation when the head is masked during forward passes on a small calibration set of reasoning problems. These causal scores are then converted into weight-level importance values for the corresponding projection matrices. Unlike magnitude-only or activation-based criteria, CAP's interventional measurement directly captures each head's functional contribution, yielding relative accuracy gains of up to 61% over Wanda on ARC-Challenge at 20% sparsity. We evaluate CAP on GSM8K, StrategyQA, and ARC-Challenge using Llama-3-8B-Instruct and Mistral-7B-Instruct at 10%, 20%, and 50% sparsity. At moderate sparsity (10-20%), CAP improves over Wanda in most model-benchmark configurations. with especially large gains on ARC-Challenge for Llama-3. Our results suggest that attention-head-level causal attribution can better preserve reasoning performance on downstream benchmarks than correlational pruning criteria at equivalent sparsity, while remaining limited by coarse MLP attribution at 50% sparsity.

Comments: Accepted at the ICLR 2026 Workshop on LLM Reasoning. 13 pages, 2 figures

Subjects:

Computation and Language (cs.CL)

Cite as: arXiv:2606.19350 [cs.CL]

(or arXiv:2606.19350v1 [cs.CL] for this version)

https://doi.org/10.48550/arXiv.2606.19350

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Yuhao Ge [view email] [v1] Mon, 27 Apr 2026 01:44:10 UTC (89 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled Pruning via Causal Attribution Preserves Reasoning Performance in Large Language Models, by Amogh Sheth and 4 other authors

View PDF

HTML (experimental)

TeX Source

view license

Current browse context:

cs.CL

new | recent | 2026-06

Change to browse by:

cs

References & Citations

NASA ADS

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

EngineersAdvanced

Key points

  • CAP uses causal intervention to measure each attention head's functional contribution, enabling fine-grained weight pruning without training.
  • At 10-20% sparsity, CAP outperforms Wanda on multiple reasoning benchmarks, with up to 61% relative gain on ARC-Challenge with Llama-3.
  • CAP's effectiveness is limited at higher sparsity (50%) due to coarse MLP attribution.

Highlights and analysis are generated automatically and may contain errors. Check the original source.

More recent stories on this topic · Models