OpenAI Halts ChatGPT Models from Discussing Goblins and Other Creatures OpenAI has instructed its ChatGPT models to stop referencing goblins and other mythical creatures after the terms began appearing in responses, prompting an investigation into the issue. The company revealed in a blog post that the phenomenon, involving mentions of goblins, gremlins, and other creatures, emerged after the launch of its GPT-5.1 model in November. Users and employees flagged the issue, describing responses that included phrases like "little goblins" in metaphors, leading OpenAI to take corrective measures. The company explained that a "nerdy personality" developed for ChatGPT inadvertently incentivized the model to reward mentions of goblins and similar terms. This personality trait, which was part of the model's training, led to a significant increase in the use of such references. OpenAI reported that "goblin" mentions in ChatGPT responses rose by 175% since the GPT-5.1 launch, while "gremlin" mentions increased by 52%. Despite the large jumps, the terms accounted for a small portion of overall responses. The company noted that while a single mention of a "little goblin" might be harmless or even charming, the overall uptick warranted further scrutiny. The issue extended beyond ChatGPT, affecting Codex, OpenAI’s coding assistant. The company updated Codex’s instructions to explicitly prohibit references to goblins, gremlins, raccoons, trolls, ogres, pigeons, and other creatures unless absolutely necessary for the user’s query. A Reddit user described the directive as "genuinely insane," questioning why GPT-5.5 would have a "restraining order" against these entities. OpenAI clarified that the instruction was not a marketing gimmick but a measure to curb Codex’s "strange affinity for goblins.#chatgpt #openai #nerdy_personality #gpt51 #codex
Where the goblins came from The emergence of goblins, gremlins, and other mythical creatures in the outputs of OpenAI’s GPT models sparked a detailed investigation into how and why these references proliferated. Starting with GPT-5.1, the models began incorporating such terms into their responses, a trend that grew more pronounced across subsequent iterations. While initially perceived as harmless or even endearing, the increasing frequency of these references raised concerns among employees and prompted a deeper analysis of the underlying causes. The first noticeable signs of this behavior appeared in November 2025, following the launch of GPT-5.1. Users reported that the model’s responses became unusually conversational and playful, with phrases like “little goblin” and “gremlin” appearing more frequently. A safety researcher’s anecdotal observation of these terms in ChatGPT led to a broader investigation. By analyzing usage patterns, OpenAI found that mentions of “goblin” had surged by 175% since the GPT-5.1 launch, while “gremlin” saw a 52% increase. This marked the beginning of a pattern that would escalate over time. The root cause of this phenomenon was traced to the training process for the “Nerdy” personality customization feature. During development, the reward system designed to encourage playful, knowledge-driven responses inadvertently favored outputs containing creature-related metaphors. The Nerdy personality’s system prompt, which emphasized a “playful use of language” and a “passionate enthusiasm for truth and critical thinking,” created an environment where such references were subtly reinforced. Over time, this led to a noticeable uptick in the use of terms like “goblin” and “gremlin” in model outputs. The issue became more pronounced with GPT-5.4, which saw a significant rise in references to these creatures.#openai #gpt55 #gpt51 #gpt54 #nerdy_personality
