Papers archive
Papers archive
A complete archive of Aman Sachan’s research papers and manuscript versions.
Browse the archive Use these pages when you want a cleaner directory view than the homepage.
⌕
Paper New LLM Foundry
LLM Foundry: a modular framework for orchestration, memory, compression, tools, and evaluation
A long-form draft for the repository. It formalises the architecture in the repo README, usage guide, and technical paper notes.
draft paper · public demo
adapters memory evaluation
Paper Featured v3 — 32-page manuscript
KVQuant: Adaptive Long-Context KV-Cache Compression for Memory-Bounded LLM Inference
The most complete version: rate-distortion framing, sink-aware allocation, head/layer coupling, benchmark plan, failure analysis, and a full appendixed derivation.
submission-ready preprint
LLM inference quantisation memory systems
Paper v2 — extended draft
KVQuant: Adaptive Long-Context KV-Cache Compression for Memory-Bounded LLM Inference
A strong intermediate draft with an expanded mathematical treatment and a more detailed systems analysis.
research manuscript
draft systems compression
Paper deep draft
KVQuant: Adaptive Long-Context KV-Cache Compression for Memory-Bounded LLM Inference
An earlier deep write-up, useful for tracing how the mathematics and implementation story evolved.
technical report
technical report implementation
Paper concise preprint
KVQuant: Adaptive Long-Context KV-Cache Compression for Memory-Bounded LLM Inference
A shorter, crisper manuscript if you want the idea in a more compact form.
shorter paper version
preprint compact paper
No matches. Try a different keyword.
