Does ChatGPT’s one-line style come from chat data or product tuning?
arsn202 · reddit · 2026-07-25
The post asks whether ChatGPT’s habit of breaking explanations into short one-liners may come from training on chat transcripts, including private conversations, or whether OpenAI intentionally tuned the response format.
The author notes that ChatGPT often answers with a summary, then a divider/title, then a series of short lines instead of a single coherent paragraph. They wonder whether this mirrors how people text friends, or whether it is a deliberate product choice to make responses feel more readable.
The thread is essentially a question about whether the model learned the style from data or from alignment and formatting decisions, and why OpenAI would prefer that style if it feels unnatural in text.
More from Models
- Opus 5 beats Fable 5 on six agentic benchmarks, suggesting a split-role setup — daniel_mac8 · 2026-07-25
- A model eval can matter more for who finishes second than for who wins — ziv_ravid · 2026-07-25
- Artificial Analysis chart says Claude Opus 5 costs about 2× GPT 5.6 Sol per task — soumitrashukla9 · 2026-07-25
- Claude Opus 5 is now available in Hermes Agent via Nous Portal and OpenRouter — NousResearch · 2026-07-25
- An Opus 5 meme turns “a rose” into a tensor joke — liminal_bardo · 2026-07-25
- Early Opus 5 access suggests the model is fast, but too eager at high reasoning — alliekmiller · 2026-07-25