Benchmark Radar™
Star —

Inverse IFEval

Explore score history ↗

Tests ability to override training biases and follow counterintuitive instructions; complementary to IFEval.

Released
2025-09-04
Evaluates
instruction following
Openness
unknown
Importer
Model reports

Inverse IFEval paper

Inverse IFEval results and reported scores

Reported benchmark scores · Model reports

Reported scores: 3

ModelOrganizationReported valueReportedEvidence
Seed2.0 LiteByteDance77.12026-02-14evidence
Seed2.0 MiniByteDance69.32026-02-14evidence
Seed2.0 ProByteDance78.92026-02-14evidence

Scores are partitioned by the source that reported them and are never merged into a single cross-source ranking, because the sources measure different things and say so.

Which sources cite Inverse IFEval?

Related instruction following benchmarks