Share your thoughts, 1 month free Claude Pro on usSee more

Home/Benchmarks

Reasoning on ANLI R3

48.25Accuracy

GANPO

Updated 1mo ago

Evaluation Results

Method	Links
GANPO 2026.01		48.25
DPO 2026.01		47.92
Base 2026.01		47.75

SOTA Paper

GANPO

Latent Adversarial Regularization for Offline Preference Optimization

Dataset

ANLI

Follow for update

@wizwand_team Discord

Related Benchmarks

Preference Alignment on AlpacaEval weighted gpt4 turbo 2.0

© 2026 wizwand

Blog Contact Changelog Swarm

Privacy Policy Terms of Service FAQs Swarm Docs