Pith. sign in

REVIEW 7 cited by

Dissecting Language Models: Machine Unlearning via Selective Pruning

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2403.01267 v2 pith:2QLLP4UQ submitted 2024-03-02 cs.LG cs.CL

Nicholas Pochinkov , Nandi Schoots This is my paper · ORCID
classification cs.LGcs.CL
keywords llmsneuronsmethodlanguagemachinemodelspruningselective
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Understanding and shaping the behaviour of Large Language Models (LLMs) is increasingly important as applications become more powerful and more frequently adopted. This paper introduces a machine unlearning method specifically designed for LLMs. We introduce a selective pruning method for LLMs that removes neurons based on their relative importance on a targeted capability compared to overall network performance. This approach is a compute- and data-efficient method for identifying and removing neurons that enable specific behaviours. Our findings reveal that both feed-forward and attention neurons in LLMs are specialized; that is, for specific tasks, certain neurons are more crucial than others. Code from all experiments is available at https://github.com/nickypro/selective-pruning

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 7 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. ExAnte: A Benchmark for Ex-Ante Inference in Large Language Models

    cs.LG 2025-05 conditional novelty 7.0 of 10

    Models leak future knowledge despite explicit temporal cutoffs, as quantified by the ExAnte benchmark across four tasks.

  2. SECNEURON: Reliable and Flexible Abuse Control in Local LLMs via Hybrid Neuron Encryption

    cs.CR 2025-06 conditional novelty 6.0 of 10

    SECNEURON uses per-neuron AES encryption plus attribute-based key management so a locally deployed LLM can be selectively decrypted to allow only authorized tasks and prune unauthorized capabilities.

  3. Model Unlearning via Sparse Autoencoder Subspace Guided Projections

    cs.CL 2025-05 conditional novelty 6.0 of 10

    SSPU uses SAE-derived subspaces to guide weight updates, lowering WMDP-Cyber accuracy by 3.22% more than RMU while largely preserving MMLU, TruthfulQA, and GSM8K performance.

  4. Module-Aware Parameter-Efficient Machine Unlearning on Transformers

    cs.LG 2025-08 conditional novelty 5.0 of 10

    MAPE-Unlearn uses Fisher-information-based scores and greedy search to select important heads and filters, then applies sparse unlearning updates, claiming improved efficacy-fidelity trade-offs on Transformers.

  5. Verifiable Unlearning on Edge

    cs.LG 2025-06 reject novelty 5.0 of 10

    A pruning-plus-OBS unlearning method is wrapped in a proposed zk-SNARK verification protocol, but no proof-generation evaluation is provided and the single experiment is self-referential.

  6. SoK: Machine Unlearning for Large Language Models

    cs.LG 2025-06 conditional novelty 5.0 of 10

    A new taxonomy for LLM unlearning distinguishes removal-intended from suppression-intended methods, and argues that gradient ascent methods functionally behave like suppression.

  7. Machine Unlearning for Robust DNNs: Attribution-Guided Partitioning and Neuron Pruning in Noisy Environments

    cs.LG 2025-06 reject novelty 4.0 of 10

    The paper proposes attribution-guided data partitioning plus regression-based neuron pruning and fine-tuning for noisy training data, but the headline label-noise result is contradicted by the feature-noise-only experiments.

Pith tools