Newest papers

When Thoughts Become Memory: Unifying In-Context Learning and Fine-Tuning

Leon Chlon, Julie Huang · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.29419 · cited by 0 works (OpenAlex) · open-access PDF

Korean

우리는이 카메라 네트워크 컨텍스트는 프레드 트레이닝 된 모델의 컨텍스트 또는 그 무게에 대한 보이지 않는 데이터가 프레드 트레이닝 된 모델의 컨텍스트에 속하는지 여부에 대한 인증 된 규칙을 제공하는 최초의 카메라 컨텍스트입니다. 그 값, 체크 아웃 쿼리에 저장된 코드 길이는 예상에 저장된 비트와 동등합니다. 모든 인증서는 훈련 된 변압기, TabPFN, 직무 네트워크, 스테레오 네트워크에 대한 디코딩 손실을 감소합니다. 법의 천장은 하나의 측정에서 선택을 인증하고, 언제든지 유효한 테스트는 체크 아웃 쿼리에서 훈련 된 것으로 만듭니다. 모든 인증서는 훈련 된 변

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

Abstract

We are the first to give a certified rule for whether unseen data belong in a pretrained model's context or its weights, built on an exact law for their information value. That value, the code length saved on held-out queries, equals in expectation the bits stored about the query, plus a prior repair, minus a decoding loss, for any model, architecture, dataset and task. The law's ceilings certify the choice from one measurement, and an anytime-valid test makes it from held-out queries, with error controlled whenever it stops. Every certificate the ceilings issue on a trained transformer, TabPFN, an occupancy network and a stereo network is correct. The test picks TabPFN over XGBoost and a trained surrogate over an in-context learner on airfoil pressure, calls the better of GPT-2's cache and fine-tune in 61,948 of 64,000 query streams over 32 comparisons, and errs in at most 3.3% of null streams. The law also diagnoses when learning in context helps and when it does not. Deleting part of a context is held-out validation of that part in any model, and adding parts one at a time traces the learning curve in context: one patch that sees a shape feature cuts an occupancy network's code length there a hundredfold, while eight that miss it nearly double it, and one camera that sees the surface beats four in a stereo network. Ranked by this label-free valuation, a kriging learner's three best support cars beat four drawn at random, and on a yes-or-no claim the valuation recovers a published evidence adjudicator, with AUROC 0.91.

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

Topics: Computer science, Certificate, Artificial intelligence, Context (archaeology), Code (set theory), Cache

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