List your company

Logical Reasoning on KK 4–5 ppl.

60.3AccuracyOracle Verifier
-0.33215.40931.1546.891May 31, 2026

Evaluation results

MethodLinks
Oracle VerifierModel Size=12B2026.0560.3
RevisionSEModel Size=12B2026.0558.7
SimpleSEModel Size=12B, Verifier threshold (τ)=0.72026.0556.3
SimpleSEModel Size=12B, Verifier threshold (τ)=0.62026.0555
SimpleSEModel Size=12B, Verifier threshold (τ)=0.82026.0553.9
SimpleSEModel Size=12B, Verifier threshold (τ)=0.52026.0553.7
Oracle VerifierModel Size=4B2026.0552.7
gemma-3-12b-itModel Size=12B2026.0551.9
RevisionSEModel Size=4B2026.0546.4
SimpleSEModel Size=4B, Verifier threshold (τ)=0.62026.0545.4
SimpleSEModel Size=4B, Verifier threshold (τ)=0.82026.0544.6
SimpleSEModel Size=4B, Verifier threshold (τ)=0.72026.0543.9
SimpleSEModel Size=4B, Verifier threshold (τ)=0.52026.0539.1
gemma-3-4b-itModel Size=4B2026.0531
Oracle VerifierModel Size=1B2026.0510
gemma-3-1b-itModel Size=1B2026.054.9
RevisionSEModel Size=1B2026.054.7
SimpleSEModel Size=1B, Verifier threshold (τ)=0.82026.054.5
SimpleSEModel Size=1B, Verifier threshold (τ)=0.52026.053.5
SimpleSEModel Size=1B, Verifier threshold (τ)=0.72026.053.3
SimpleSEModel Size=1B, Verifier threshold (τ)=0.62026.052