JSON evaluation cases
evalbench.nsisong.com
evalbench
A repeatable evaluation platform for prompts, expected behavior, regressions, safety cases, latency, and cost.
JSON run outputs
expected-marker checks
latency and cost totals
Markdown reports
evalbench.nsisong.com
A repeatable evaluation platform for prompts, expected behavior, regressions, safety cases, latency, and cost.
JSON evaluation cases
JSON run outputs
expected-marker checks
latency and cost totals
Markdown reports