Prefill-Free Cross-Family KV Cache Transfer for Heterogeneous Multi-Agent LLMs Paper • 2609.32259 • Published Sep 29 • 20
Unlocking Structured Pruning for Reasoning Language Models via Attention-Preserving MLP Trimming Paper • 2610.19121 • Published Oct 13 • 6
ResidualQuant: KV Cache Quantization for Looped Transformers with 2-Bit Residuals Paper • 2610.10381 • Published Oct 7 • 9
Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs Paper • 2605.15491 • Published Jun 7 • 1
Compressed Models V1 Collection Compressed LLMs (Qwen3, Llama families). V1 (First Version) • 16 items • Updated Jul 1 • 1