Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
dni138
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
2 ms
·
1.
▲
by
dni138
8mo ago
Appreciate the comment here! And I agree. Generative models, for the most part, are very much controlled by their context window. So if you provide it a prompt, such as the prompt injection attack you seem to describe, yes the LLM will perf