← Search

Suppakit Waiwitlikhit

2 accepted papers

2026

Breaking Barriers: Do Reinforcement Fine-tuning Gains Transfer To Unseen Domains?

ICLR 2026poster

Reinforcement post training (RPT) has recently shown promise in improving the reasoning abilities of large language models (LLMs). However, it remains unclear how well these improvements generalize to new domains, as prior work evaluates RPT models on data from the same domains used for fine-tuning.…

Cited by 0SourceScholar
2024

Trustless Audits without Revealing Data or Models

ICML 2024poster

There is an increasing conflict between business incentives to hide models and data as trade secrets, and the societal need for algorithmic transparency. For example, a rightsholder who currently wishes to know whether their copyrighted works have been used during training must convince the model pr…

Cited by 8SourcePDFScholar