Suhaas Bhat — accepted AI-conference papers

2026

Building Reliable Long-Form Generation via Hallucination Rejection Sampling

Lin Li, Georgia Channing, Suhaas Bhat, Gabriel Jones, Yarin Gal

ICML 2026poster

Large language models (LLMs) have achieved remarkable progress in open-ended text generation, yet they remain prone to hallucinating incorrect or unsupported content, which undermines their reliability. This issue is exacerbated in long-form generation due to hallucination snowballing, a phenomenon …

Cited by 0 Source ↗Scholar ↗

2026

Rubric Curriculum RL: Exploiting the Generation-Verification Gap in Creative Writing

Tejas Krishnan, Sumeet Motwani, Charles London, Suhaas Bhat, Huitian Jiao, Phil Torr +5

ICML 2026poster

Reinforcement learning with verifiable rewards (RLVR) on foundation models has led to significant improvements in math and code generation. Extending these gains to open-ended domains remains challenging: ground-truth verification is unavailable, human annotation is expensive, and learnt reward mode…

Cited by 0 Source ↗Scholar ↗