Screen capture from OpenAI documentation platform
Loading data from Talking to Chatbots Dataset reddgr/talking-to-chatbots-chats …
😎 The screen capture is from https://platform.openai.com/docs

🤖 Screen capture from OpenAI’s platform documentation showing a section titled “How reasoning works.” It describes how models use reasoning tokens to process prompts, breaking down the problem before generating a response. The text explains that reasoning tokens are discarded after producing output tokens. Below the explanation, a diagram illustrates a multi-step conversation between a user and an assistant across three turns. Each turn consists of “Input,” “Reasoning,” and “Output.” Arrows connect the turns, showing how output tokens are carried over, but reasoning tokens are discarded. At the bottom, a dotted line labeled “Context window 128k tokens” marks the truncation of output tokens.
[Alt text by ALT Text Artist GPT]
Hugging Face Dataset Metrics
All the conversation prompts, responses, and metrics are available to download and explore on Hugging Face dataset reddgr/talking-to-chatbots-chats: