xwn

4 ms·

1. 1.▲
   by xwn 2y ago

   Check the last entry in the FAQ source
2. 2.▲
   by xwn 2y ago

   Static prompts are a downside of using academic research in a tool like this. Two notes: * ineffective prompts come out of garak and new prompts come in to garak, so eval scores always drop over time on a static target * there are more and
3. 3.▲
   by xwn 2y ago

   The proof has been in the pudding
4. 4.▲
   by xwn 2y ago

   Thanks! Wrote it loooong before it was a corporate tool and was only a labor of love. Now it's both
5. 5.▲
   by xwn 3y ago

   free *direction for big tech, who aren't particularly far forward on llmsec anyway
6. 6.▲
   by xwn 3y ago

   English verbs nouns all the time
7. 7.▲
   FakeToxicityPrompts: Automatic Red Teaming(interhumanagreement.substack.com)

   57 points by xwn 3y ago|55 comments
8. 8.▲
   People keep referring to LLMs as demons. Why?(interhumanagreement.substack.com)

   2 points by xwn 4y ago|0 comments
9. 9.▲
   by xwn 4y ago

   I don't know, without enumerating risks to check, there's little basis for doing due diligence and quelling investors. This massively-cited paper gave a good point of departure for establishing rigorous use of LLMs in the real wor