RAG, Fine-Tuning & ContextAshwin Gerard Colaco, Nada Lahjouji

What to Keep, What to Forget published

What to Keep, What to Forget published — frames KV cache management, prompt pruning and agent memory consolidation together as a rate-distortion problem

Why it matters

Organizes previously scattered research on context compression into a single theoretical framework

Source — Primary sources first; secondary reports are labelled “Press”

Paperarxiv.orghttps://arxiv.org/abs/2607.08032

Added in Week 35, 2026 · See that week

Same week

Thinking Machines Lab posts The Future Worth Building Is HumanOfficialNYT and Daily News seek sanctions, alleging OpenAI hid evidencePressOllama raises $89 million; 8.9 million developersOfficialOpenClaw Foundation launches as a 501(c)(3) nonprofit, chaired by Dave MorinOfficial

Nearby on the same axis

Diagnosing and Mitigating Context Rot in Long-horizon Search publishedPaperMartin and Roger publish Classifier Context RotPaperKDDI shares how it optimized its Buffmee RAG app on Google CloudOfficialSun et al. present AutoSearchPaper