Latent Compaction with Test-Time Training
We introduce Latent Compaction, which learns a compact document-specific cache through test-time training.
ProofPaper ↗
Key points
- Long contexts enlarge key--value (KV) caches, increasing memory use and attention costs during generation.
- Compressing these caches must preserve information for unseen questions.
- We initialize cache entries using reconstruction importance, teacher attention, and key-space diversity.
- These results show that treating the KV cache as a neural network and optimizing it with test-time training can preserve retrieval accuracy under aggressive compression.
Sources (2)
- [1]Latent Compaction with Test-Time TrainingarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 05:57 PM
We introduce **Latent Compaction**, which learns a compact document-specific cache through test-time training.
Long contexts enlarge key--value (KV) caches, increasing memory use and attention costs during generation.
- [2]Latent Compaction with Test-Time TrainingHugging Face Daily Papers · Oct 14, 12:00 AM · same content
Extractive summary: sentences quoted from the sources.
Before this
- Oct 13, 2026Run fast, private agents on-device with LFM2.5-24B-A2B
- Oct 13, 2026Gemini 4 Argon: Built to solve your hardest problems
- Oct 13, 2026LongForcing: Scaling Temporal Manifold Matching for Long-Horizon World Rollout
- Oct 13, 2026LinMorph: Learning to Morph Weights for Efficient Video DiT Linearization
- Oct 13, 2026Fish Audio Drama 3 Technical Report
- Oct 12, 2026MLAfication: Efficiently Enabling High-Fidelity MLA for Modern LLMs