Tejal Patwardhan, head of the frontier evals team, spoke with Andrew Mayne about how traditional benchmarks are becoming saturated or can be manipulated, so it is important to apply new, more comprehensive methods to measure and forecast models' progress. The discussion urges expanding the evaluation criteria for models.
New evaluation methods are needed to reliably measure models
Tejal Patwardhan, head of the frontier evals team, spoke with Andrew Mayne about how traditional benchmarks are becoming saturated or can be manipulated, so it is important to apply new, more comprehensive methods to measure and forecast models' progress.
