UiPath

2 primary articles · 0 secondary mentions

Primary coverage

Major research UiPath

Latent Adversarial Detection: Adaptive Probing of LLM Activations for Multi-Turn Attack Detection

Problem This paper addresses the gap in the detection of multi-turn prompt injection attacks on large language models (LLMs), specifically focusing on covert attacks where individual turns appear benign. Existing...

arXiv cs.AI Apr 30, 2026 alignment safety 459w arXiv code Prashant Kulkarni

Notable research UiPath

Auto-FlexSwitch: Efficient Dynamic Model Merging via Learnable Task Vector Compression

Problem This paper addresses the challenge of efficient dynamic model merging for multi-task adaptation, specifically focusing on the storage overhead associated with maintaining independent parameters for each task. Existing dynamic...

arXiv cs.LG Apr 30, 2026 efficiency inference 474w arXiv code Junqi Gao +5