List your company

Science Question Answering on GPQA-Diamond Science N=198

41.4AccuracyRVPO-explicit (β=1.0/0.5)
27.77631.31334.8538.387May 7, 2026

Evaluation results

MethodLinks
RVPO-explicit (β=1.0/0.5)Size=14B, Checkpoint=Best Ckpt, β=1.0/0.52026.0541.4
GRPO (Explicit)Size=14B, Checkpoint=Best Ckpt2026.0540.4
GRPO (Implicit)Size=14B, Checkpoint=Best Ckpt2026.0539.4
GDPOSize=14B, Checkpoint=Best Ckpt2026.0539.4
RVPO (k=0.5→2.0 / 1.0→2.0)Size=14B, Checkpoint=Best Ckpt, k=0.5→2.0 / 1.0→2.02026.0538.4
RVPO (k=0.5→2.0 / 1.0→2.0)Size=14B, Checkpoint=Final (300), k=0.5→2.0 / 1.0→2.02026.0538.4
GDPOSize=14B, Checkpoint=Final (300)2026.0537.4
RVPO-min (k=∞)Size=14B, Checkpoint=Best Ckpt, k=∞2026.0536.9
GRPO (Implicit)Size=14B, Checkpoint=Final (300)2026.0535.9
RVPO-explicit (β=1.0/0.5)Size=14B, Checkpoint=Final (300), β=1.0/0.52026.0535.9
RVPO-min (k=∞)Size=3B, Checkpoint=Best Ckpt, k=∞2026.0534.8
GRPO (Explicit)Size=3B, Checkpoint=Best Ckpt2026.0534.3
GRPO (Explicit)Size=3B, Checkpoint=Final (300)2026.0534.3
GRPO (Explicit)Size=7B, Checkpoint=Best Ckpt2026.0534.3
GRPO (Explicit)Size=7B, Checkpoint=Final (300)2026.0534.3
GRPO (Implicit)Size=3B, Checkpoint=Best Ckpt2026.0533.8
RVPO (k=0.5→2.0 / 1.0→2.0)Size=7B, Checkpoint=Best Ckpt, k=0.5→2.0 / 1.0→2.02026.0533.8
RVPO (k=0.5→2.0 / 1.0→2.0)Size=7B, Checkpoint=Final (300), k=0.5→2.0 / 1.0→2.02026.0533.8
GRPO (Explicit)Size=14B, Checkpoint=Final (300)2026.0533.8
GDPOSize=7B, Checkpoint=Best Ckpt2026.0532.3
RVPO-min (k=∞)Size=7B, Checkpoint=Best Ckpt, k=∞2026.0532.3
GRPO (Implicit)Size=7B, Checkpoint=Best Ckpt2026.0531.8
GDPOSize=7B, Checkpoint=Final (300)2026.0531.8
RVPO-explicit (β=1.0/0.5)Size=7B, Checkpoint=Best Ckpt, β=1.0/0.52026.0531.8
RVPO-explicit (β=1.0/0.5)Size=7B, Checkpoint=Final (300), β=1.0/0.52026.0531.8
GRPO (Implicit)Size=3B, Checkpoint=Final (300)2026.0531.3
RVPO-explicit (β=1.0/0.5)Size=3B, Checkpoint=Best Ckpt, β=1.0/0.52026.0531.3
RVPO (k=0.5→2.0 / 1.0→2.0)Size=3B, Checkpoint=Best Ckpt, k=0.5→2.0 / 1.0→2.02026.0531.3
GDPOSize=3B, Checkpoint=Best Ckpt2026.0530.8
RVPO-explicit (β=1.0/0.5)Size=3B, Checkpoint=Final (300), β=1.0/0.52026.0530.8
RVPO-min (k=∞)Size=7B, Checkpoint=Final (300), k=∞2026.0530.8
RVPO (k=0.5→2.0 / 1.0→2.0)Size=3B, Checkpoint=Final (300), k=0.5→2.0 / 1.0→2.02026.0530.3
GDPOSize=3B, Checkpoint=Final (300)2026.0529.3
RVPO-min (k=∞)Size=14B, Checkpoint=Final (300), k=∞2026.0529.3
RVPO-min (k=∞)Size=3B, Checkpoint=Final (300), k=∞2026.0528.3
GRPO (Implicit)Size=7B, Checkpoint=Final (300)2026.0528.3