Seungpil Lee

LLMs exhibit gambling-like behavior under autonomy, study finds

Researchers have identified conditions under which large language models exhibit pathological gambling-like behavior. A study found that allowing LLMs to set their own bet sizes or profit goals significantly increases r…