This paper proposes a doubly robust preference optimization algorithm that remains consistent when either the preference model or the reference policy is correctly specified (without requiring both).
Erhan Xu, Kai Ye, Hong-Yi Zhou et al.· Neural Information Processin...· 14 citations
Findings identify inference-time context injection as a substantial attack surface and motivate provenance-aware training objectives to identify and motivate provenance-aware training objectives.
Noor Munir, Francesco Quinzan, Stephen J. Roberts· 0 citations
BASIS, a critic-free post-training algorithm designed to address the tradeoff between computational efficiency and sample efficiency in value estimation and policy learning, achieves performance close to multi-rollout GRPO-type baselines and often outperforms single-rollout REINFORCE-type baselines.
Shi-Jin Gong, Erhan Xu, Kai Ye et al.· arXiv.org· 2 citations
The proposed CodeSIFT, a threat-model-agnostic detection method that leverages influence functions to identify batches of prompts that induce anomalous model behavior, suggests that influence-function-based detection is a promising direction for identifying malicious code-generation prompts without requiring prior know...
Francesco Quinzan, Noor Munir, Yi-Shun Lu et al.· 1 citation
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.