Alex Strick van Linschoten

Historian turned AI/ML engineer, at ZenML since 2021. Writing about agents , evals and RL , and about Afghanistan , languages and books .

Visit mlops.systems → On Hacker News Jun 2024 – Jul 2024

Finetunes of Mistral, Llama3 and Solar LLMs are more accurate for my test data than OpenAI’s models.

I summarise the kinds of evaluations that are needed for a structured data generation task.