← Search

Varun Prashant Gangal

2 accepted papers

2026

Position: We Need A Unified Definition of Hallucination (It’s The World Model, Stupid!)

ICML 2026poster

Despite numerous attempts at mitigation since the inception of language models, hallucinations remain a persistent problem even in today's frontier LLMs. Why is this? We review existing definitions of hallucination and fold them into a single, unified definition wherein prior definitions are subsume…

Cited by 0SourceScholar
2025

Sparse Rewards Can Self-Train Dialogue Agents

ACL 2025finding

Recent advancements in state-of-the-art (SOTA) Large Language Model (LLM) agents, especially in multi-turn dialogue tasks, have been primarily driven by supervised fine-tuning and high-quality human feedback. However, as base LLM models continue to improve, acquiring meaningful human feedback has be…