github.com
GitHub - achalagrawal/dhee-eval: Reusable MCP tool-calling eval harness: sweep candidate LLMs across assistant profiles, score with an LLM judge. Built to pick models for the Madhyasth Darshan study a...
Reusable MCP tool-calling eval harness: sweep candidate LLMs across assistant profiles, score with an LLM judge. Built to pick models for the Madhyasth Darshan study assistant. - achalagrawal/dhee-...