§ feed · storyline
What do LLMs think when you don't tell them what to think about?
New research finds LLM families have distinct default knowledge priors, with GPT favouring code and math, Llama narratives, DeepSeek religious content, and Qwen exam questions when given no prompt.
What do language models generate when you don't tell them what to generate? New research reveals that LLM families have distinct 'knowledge priors'—GPT models default to code and math, Llama favors narratives, DeepSeek generates religious content, and Qwen outputs exam questions.
§ sources1 publication · timeline below