In-Place TTT adapts LLM MLP projection matrices at test time with a next-token-aligned objective and chunk-wise updates, enabling better long-context performance as a drop-in enhancement.
Memoryllm: Towards self-updatable large language models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
In-Place Test-Time Training
In-Place TTT adapts LLM MLP projection matrices at test time with a next-token-aligned objective and chunk-wise updates, enabling better long-context performance as a drop-in enhancement.