AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Positional Cognitive Specialization: Where Do LLMs Learn to Comprehend and Speak Your Language?

Luis Frentzen Salim, Lun-Wei Ku, Hsing-Kuo Kenneth Pao

AAAI 2026 · Large language models › Language, safety and interpretability · approach: Multilingual LLM data and culture · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
train only the 25% outermost layers (CogSym) instead of all layersLlama 2 7B · Javanese/Sundanese downstream taskstask performance: within 2-3% of full finetuningNo clear effect
train rear layers only instead of front layers only (ablation)Llama 2 7B · word translation to Javanese/SundaneseaccuracyImproved
train front layers only instead of rear layers only (ablation)Llama 2 7B · word translation from Javanese/SundaneseaccuracyImproved
train both front and rear regions together instead of one region with same per-region budget (ablation)Llama 2 7B · word translationaccuracyImproved
move trained layers inward from the model extremities (k=8) (ablation)Llama 2 7B · word translation (Javanese)accuracyHurt
select layers by FIM instead of CogSym positional heuristic (ablation)Llama 2 7B · word translation (Javanese)accuracyNo clear effect
select layers by LSN count or AlphaLoRA instead of CogSym (ablation)Llama 2 7B · word translation (Javanese)accuracyHurt
apply CogSym with LoRA instead of full finetuningLlama 2 7B · word translation (Javanese)accuracyNo clear effect
train regions sequentially (rear first) instead of simultaneously (ablation)Llama 2 7B · translation toward unknown languageaccuracy: almost 10% degradationHurt
train regions sequentially (front first) instead of simultaneously (ablation)Llama 2 7B · translation toward unknown languageaccuracyNo clear effect
train more than ten layers in one specialized region (ablation)Llama 2 7B · region-specific tasksaccuracyNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy