Pinned Loading

1. nanoreasoner nanoreasoner Public

   Knowledge distillation fails at 158x compression: a systematic negative result with statistical rigor

   Python 2

2. distillation-reward-audit distillation-reward-audit Public

   Post-hoc audit of a 22.3σ val_bpb false-positive in a 19M-parameter knowledge distillation experiment. Dual-dimension attribution: token-level gradient + sample-level pass@k.

   Python 1

3. harness-as-governance harness-as-governance Public

   Harness is governance, not prompt engineering

   1

4. DeepSeek-V3.2-Exp DeepSeek-V3.2-Exp Public

   Forked from deepseek-ai/DeepSeek-V3.2-Exp

   Python 1

5. vllm vllm Public

   Forked from vllm-project/vllm

   A high-throughput and memory-efficient inference and serving engine for LLMs

   Python 1

6. FlashMLA FlashMLA Public

   Forked from deepseek-ai/FlashMLA

   FlashMLA: Efficient Multi-head Latent Attention Kernels

   C++ 1