dhavalt

3 ms·

1. 1.▲
   by dhavalt 3mo ago

   Working on a desktop application for llm evaluations.
2. 2.▲
   Show HN: AptSelect – A local LLM client for parallel testing and evaluation(aptselect.com)

   3 points by dhavalt 3mo ago|0 comments
3. 3.▲
   by dhavalt 4mo ago

   Always curious about which llms perform best in specific scenarios, so I built a local desktop app to benchmark and evaluate prompts and llms side by side.
4. 4.▲
   by dhavalt 6mo ago

   Love seeing more local desktop apps in these threads, especially ones tackling local network sync like Kitted. I'm a solo dev also in the trenches building a local desktop tool right now, so I know how fun (and painful) those architect
5. 5.▲
   Show HN: AptSelect – A local desktop app to test LLMs side-by-side(aptselect.com)

   2 points by dhavalt 6mo ago|0 comments
6. 6.▲
   by dhavalt 9mo ago

   While debugging a crash or incident, I used to get lost in the specific line of code and trying to fixing symptoms. TLA+ gave me new prespective towards debugging, and started treating system crashes as a state problems rather than just cod
7. 7.▲
   Show HN: Latent Logbook – Prompt Puzzles Built with Hapi.js and SQLite(latentlogbook.com)

   1 points by dhavalt 10mo ago|0 comments
8. 8.▲
   Vending-Bench: The Simulation Exposing LLMs' Long-Term Focus Problem(blog.dhavaltanna.com)

   3 points by dhavalt 1y ago|0 comments