Chapter 5: Functions
The BrokenMath benchmark (NeurIPS 2025 Math-AI Workshop) tested this in formal reasoning across 504 samples. Even GPT-5 produced sycophantic “proofs” of false theorems 29% of the time when the user implied the statement was true. The model generates a convincing but false proof because the user signaled that the conclusion should be positive. GPT-5 is not an early model. It’s also the least sycophantic in the BrokenMath table. The problem is structural to RLHF: preference data contains an agreement bias. Reward models learn to score agreeable outputs higher, and optimization widens the gap. Base models before RLHF were reported in one analysis to show no measurable sycophancy across tested sizes. Only after fine-tuning did sycophancy enter the chat. (literally)
，这一点在WhatsApp Web 網頁版登入中也有详细论述
^ See, e.g., John C.P. Goldberg & Benjamin C. Zipursky, Recognizing Wrongs 2 (2020).
Sorry, something went wrong.。关于这个话题，谷歌提供了深入分析
15 // reset to the main entry point block to keep emitting nodes into the correct conext
into your own lisp/ directory and require it. That's it.，更多细节参见whatsapp