Newest papers

Can Language Models Get "Aphasia" From Neuronal Ablation?

Pratik Worah, Randall Bateman · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.28214 · cited by 0 works (OpenAlex) · open-access PDF

Korean

그러나 신경 질환의 손실이 인지 감소의 패턴을 형성하는 방법은 여전히 잘 이해되지 않습니다. 여기서 우리는 체계적으로 신경 단위 (컴퓨팅 단위) 언어 모델에 반복적 인 단기 기억 (LSTM) 및 트랜스퍼러 아키텍처를 범위로 반복적 인 단기 감소 (LSTM)의 특징입니다. 여기서 우리는 체계적으로 신경 단위 (컴퓨팅 단위)를 반복적 인 단기 기억 (LSTM) 및 트랜스퍼러 아키텍처에 걸쳐 반복적 인 단기 감소 (LSTM) 패턴을 방해하고 보완적 인 단순한 프록시 (MLP)를 보완적 인 플라스틱성에 적용합니다. 전체 단위 폐기와 함께 LSTM는 단계적으로

Machine translation (Meta M2M100); the English original follows.

Abstract

Progressive neuronal loss is the hallmark of neurodegenerative disease, yet how loss of neurons shapes the pattern of cognitive decline remains poorly understood. Here we systematically ablate neurons (computational units) in language models spanning recurrent long short-term memory (LSTM) and Transformer architectures and apply limited fine-tuning as a simple proxy for compensatory plasticity. With whole-unit ablation, LSTMs degrade gradually with syntactic and semantic faculties declining in parallel across mild-to-moderate damage. By contrast, Transformers exhibit late threshold-like collapse under feedforward multi-layer perceptron (MLP) ablation and smoother sigmoid-like decline under attention-head ablation. Qualitatively, MLP-ablated models continue to make within-category semantic substitutions (semantic paraphasias) alongside rising non-responses, paralleling Wernicke-type aphasia, while attention-head ablation produces a trend toward loss of semantic specificity similar to the semantic variant of primary progressive aphasia (svPPA). These findings indicate that feedforward and attentional pathways have functionally distinct failure modes, suggesting that architecture-dependent failure profiles may offer new computational perspectives on aphasic syndromes and neurodegenerative diseases as an in silico model. Notably, early accuracy retention under Transformer attention-head ablation qualitatively parallels pre-symptomatic reserve before the 40--60% neuronal loss reported by Gomez-Isla et al. (1996) when clinical symptoms emerge in Alzheimer's disease, whereas the steadier LSTM decline parallels the symptomatic progression described by DeKosky and Scheff (1990) and Terry et al. (1991).

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Parallels, Aphasia, Neuroscience, Computer science, Transformer, Semantic memory

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange