amitness
Re:New tools for building agents
on Wednesday March 12, 2025 @04:52AM (#43341078)
It's not enabled by default for their function calling API. So, hallucination is possible.
You have to set 'strict' to True manually to use the same grammar-based sampling they use for structured outputs.
https://platform.openai.com/docs/guides/function-calling?api...
Re:Show HN: I made a website to semantically search ArXiv papers
on Thursday December 26, 2024 @04:08AM (#42514042)
Try this trick that I learned from Cohere:
- Fetch top 10*k (i.e. 100) results using the hamming distance
- Rerank by taking dot product between query embedding (full precision) and binary doc embeddings
- Show top-10 results after re-ranking
Re:Open source inference time compute example from HuggingFace
on Saturday December 21, 2024 @09:42AM (#42479953)
OpenAI recommends using o1 to generate the verbose plan and then chain the verbose output to a cheaper model (e.g. gpt-4o-mini) to convert it into structured data / function calls / summary etc. They call it planner-executor pattern. [1]