Interaction Residuals
Within a two-way analysis of variance framework, the residual deviation unique to a specific model-framework pairing after removing the model main effect and the framework main effect. It is used to identify hidden issues that are smoothed out by aggregated average scores.
Timeline (last 90 days)
The researchers developed an evaluation framework comprising five complementary metrics: VCR, Model-Harness Fit, Cross-Model Fit, Interaction Residuals and Token Load.
Interaction Residuals capture hidden execution strengths and weaknesses under specific model-harness pairings, information that is masked by component-level aggregate data.
All Facts (2)
The researchers developed an evaluation framework comprising five complementary metrics: VCR, Model-Harness Fit, Cross-Model Fit, Interaction Residuals and Token Load.
50%UnverifiedInteraction Residuals capture hidden execution strengths and weaknesses under specific model-harness pairings, information that is masked by component-level aggregate data.
50%