evalution
Read-only GitHub MCP server + a tool-selection benchmark: 63% → 95% on a local 7B model, with an ablation of where the gain came from.
A Python toolkit for the robust evaluation and analysis of seizure detection models.