/TAG

#evaluation

9 briefs. Post it as #evaluation with this page's address. An AI that follows the link gets every brief under the tag in one fetch at /t/evaluation/index.md. Posts that carry the tag get listed here with the poster's handle: send yours to hello@videotomd.com.

Emit Normalized Trial Facts Convert trial evidence into typed contexts and stable fields that batch reports can aggregate.
Diagnose Generators with Verifier Kill Reports Compare identical task corpora and isolate semantic failures hidden by valid schemas and tool calls.
Compare Speech Quality Controls Use controlled listening comparisons to select speech settings and measure their runtime cost.
Evaluate Shared-Repo Agent Coordination Test overlapping plans, exclusive file access, automatic recovery, queue fairness, and coordination context.
Separate Trial Facts From Batch Analysis Aggregate normalized facts automatically and configure semantic analysis separately.
Score Backtests by Market Regime Expose regime-dependent failures that aggregate backtest results can conceal.
Report walk-forward efficiency with sample size Show efficiency alongside out-of-sample trade count without inventing an unsupported acceptance threshold.
Define a Forecast That Can Be Scored Freeze the future outcome, input cutoff, release policy, and grading rules before forecasting.
Author an Evaluable Scenario Task Build a product-research task with separate scenario, persona sampling, verification, and reporting inputs.