FOSCAR: Accelerating Decode Attention over 2- and 4-Bit KV Caches through Affine Folding
cs.AI
Submitted: 2026-11-06
Updated: 2026-11-06
Code: https://github.com/FutureMLS-Lab/FOSCAR
Terminology
Sources
- CODA: Rewriting Transformer Blocks as GEMM-Epilogue Programs
- SAW-INT4: System-Aware 4-Bit KV-Cache Quantization for Real-World LLM Serving
- GEAR: An Efficient KV Cache Compression Recipe for Near-Lossless Generative Inference of LLM
- Who Says Elephants Can't Run: Bringing Large Scale MoE Models into Cloud Scale Production
- GPQA: A Graduate-Level Google-Proof Q&A Benchmark
- Kitty: Accurate and Efficient 2-bit KV Cache Quantization with Dynamic Channel-wise Precision Boost
- FlashAttention-4: Algorithm and Kernel Pipelining Co-Design for Asymmetric Hardware Scaling
- OSCAR: Offline Spectral Covariance-Aware Rotation for 2-bit KV Cache Quantization
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents