Our great sponsors
-
guidance
Discontinued A guidance language for controlling large language models. [Moved to: https://github.com/guidance-ai/guidance] (by microsoft)
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
So from OpenAIs own github https://github.com/openai/openai-cookbook/blob/main/techniques_to_improve_reliability.md.
In LLM research, prompt engineering is serious business. There are many examples were simple modifications to a prompt significantly increases accuracy on a benchmark task (e.g., chain-of-thought, in-context learning, system prompts, etc.). If you are building apps on top of the API, then you want a prompt template that will reliably do what you hope it will do across different inputs, and you want to have predictable failure modes. I suggest you work through some of the Guidance tutorials, you'll learn a bunch: https://github.com/microsoft/guidance