AI Paper Atlas

Research map / Continual learning / Continual fine-tuning of LLMs

Skill Neologisms: Towards Skill-based Continual Learning

Antonin Berthon, Nicolás Astorga, Mihaela van der Schaar

ICML 2026 · Continual learning › Continual fine-tuning of LLMs · approach: Low-rank and model-merging adaptation · 0 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
insert skill neologism tokens in prompt (vs LoRA) for composing the new skill with held-out skillsQwen2.5-0.5B on digit-sequence skills · OOD 2-combinations, SHIFTaccuracy: 0.16 -> 0.75 (LoRA vs skill neologisms)Improved
insert soft tokens inside the prompt as a neologism rather than as a prefix (prompt tuning)skill neologisms · OOD 2-combinations, INV-POLaccuracy: 0.57 -> 0.84Improved
compose independently learned SHIFT and INV-POL neologisms zero-shot instead of in-context examplesskill neologisms · digit-sequence two-skill compositionaccuracyImproved
increase neologism length beyond l=20 (ablation)skill neologisms · OOD 2-combinations with held-out ADDaccuracyHurt
use very short neologism (l=1) (ablation)skill neologisms · digit-sequence skillsaccuracyHurt
train with higher k_max (more complex compositions) in the training set (ablation)skill neologisms, INV-POL · OOD 2-skill compositionsaccuracy: .36 -> .85 (k_max 1 -> 3)Improved
train with higher k_max (ablation)skill neologisms, SHIFT · OOD 2-skill compositionsaccuracy: .46 -> .72 (k_max 1 -> 3); .72 at k_max 2 and 3Improved
initialize from mean pretrained skill embedding instead of random (ablation)skill neologisms, INV-POL · OOD 2-skill compositionsaccuracy: .63 -> .85Improved
initialize from mean pretrained skill embedding instead of random (ablation)skill neologisms, SHIFT · OOD 2-skill compositionsaccuracy: .70 -> .72No clear effect
add noise to skill labels in training data (20%+) (ablation)skill neologisms, INV-POL · held-out ADD, 2-op and 3-opaccuracyHurt
add noise to skill labels (up to about 40%) (ablation)skill neologisms, INV-POL · held-out ADD, 1-opaccuracyNo clear effect
combine independently trained skill neologisms (modus ponens + statistical syllogism)Llama3.2-3B-Instruct · Skill-Mix, both skillsaccuracyImproved

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