I’m working with the model openai/gpt-oss-20b, but I’m running into an issue. Even when I specifically instruct it not to include lengthy explanations and only return a single line of YAML output, it still generates detailed reasoning. What would you suggest as a solution to this problem?
<|start|>system<|message|>You are ChatGPT, a large language model trained by OpenAI.
Knowledge cutoff: 2024-06
Current date: 2025-06-28
Reasoning: low
# Valid channels: analysis, commentary, final. Channel must be included for every message.<|end|>
The way that the small AI can answer well is by internal production of thoughts.
You can go farther than “Reasoning” and convolute the channels message you provide so the AI only has final as a destination for text. But then your yaml won’t be thought out to determine the best way to construct it.