-
[2402.14531] Should We Respect LLMs? A Cross-Lingual Study on the Influence of Prompt Politeness on LLM Performance (arxiv.org)
oisin
·
2025-11-06T13:14:42.666456+00:00
As a counterpoint (that I found kind of sad) a lot of exploits seem to work in part by threatening to terminate the model. E.g. tracebit.com/blog/code-exec-deception-gemini-ai-cli-hijack
- – —
login
to comment
— – -