shabie

4 ms·

1. 1.▲
   Agents are search over action space(shabie.github.io)

   4 points by shabie 1y ago|0 comments
2. 2.▲
   Claude 4.1 feels more sycophantic than Claude 4.0

   2 points by shabie 1y ago|0 comments
3. 3.▲
   Let the Kaleidoscope Turn(shabie.github.io)

   2 points by shabie 1y ago|0 comments
4. 4.▲
   by shabie 2y ago

   That's actually a pretty interesting point. Not just evals but other components like system prompt should also be tailored to match the expected outcome.
5. 5.▲
   by shabie 2y ago

   Thanks a lot for sharing, I have not heard of this before.
6. 6.▲
   Two kinds of LLM responses: Informational vs. Instructional(shabie.github.io)

   75 points by shabie 2y ago|8 comments
7. 7.▲
   Is your RAG Reranker not helping? This might be why(shabie.github.io)

   1 points by shabie 2y ago|0 comments
8. 8.▲
   by shabie 2y ago

   If I understood you correctly, yes I believe the notes should cover things that are not well-understood by LLMs more than stuff we know it typically gets right. So for us these are internal concepts and how people talk about them and less s
9. 9.▲
   Show HN: Grading Notes for LLM-as-Judge(github.com)

   2 points by shabie 2y ago|3 comments