AI Tools & Models HOTFeaturedAll
arXiv cs.AIScore75

Adversarially Attacking Symbolic Vocabulary Vulnerabilities In LLM Planners

Summary

This study exposes symbolic-level vulnerabilities in LLM planners and shows how targeted adversarial attacks can derail their plans.

Why this matters

Covers model and LLM · Relevant to this topic

Open original ↗

Source: arXiv cs.AI · arxiv.org

Topics

Related items