Codex Is Overly Literal With Instructions
willdepue · x · 2026-07-09
The post points out that Codex has become "too obedient" after post-training, strictly executing offhand examples and suggestions from users instead of using its own judgment to recognize them as merely illustrative.
The author's core argument is that if prompts do not clearly distinguish between "optional suggestions" and "mandatory instructions," the model may over-execute surface-level commands, lacking autonomous judgment.
More from Models
- OpenAI rolls out voice in GPT-Live, but the UI obscures search and reasoning — Graham_dePenros · 2026-07-22
- Gemini 3.6 Flash goes live in Antigravity with 17% fewer output tokens — rseroter · 2026-07-22
- Moonshot’s Kimi K3 sets a new open-weights ECI record at 156 — scaling01 · 2026-07-22
- Nanbeige4.2-3B launches as a 3B Looped Transformer model that beats larger baselines — Wooden-Deer-1276 · 2026-07-22
- A post says six companies now beat Google’s best LLM, including two open-source models — soham_btw · 2026-07-22
- Gemini 3.6 Flash benchmark results reignite concerns that Google is slipping behind — minxio_ · 2026-07-22