LLM JSON that sometimes fails to parse: why it happens, when constrained decoding fixes it, and a validate-and-retry loop for the rest.
LLM Returns Invalid JSON Once in a While: Structured Outputs vs Validate-and-Retry
Related
Nemotron 3.5 Transcribe on one 32 GB Blackwell card: the prompt is your API, so test it
One deployment returns verbatim, clean or speaker-tagged text by prompt: 3 minutes in 4.4 s on 32 GB. The clean prompt still adds speaker tags.
Our customers had already told us what Siri should say
Thousands of apps run on our engine: a radio station, a sports club, a local paper, a pastry shop,...
The AI agent wasn't confused. It was drowning in 200 tools.
How I cut wrong answers and token usage by giving the model 5 tools instead of 200. The ticket,...