Five LLMs Show Strong Female Protagonist Bias
TomLikesRobots · x · 2026-07-18
The author shares results from a name slop blocklist test: across 1,000 story beginnings generated by 5 LLMs, 76% of protagonists were female.
A more granular observation from the charts reveals that only 35% of secondary characters were female, with the most common character pairing being 'female protagonist + male side character' at 57%. The author also notes that the name Elara, generated by Gemini, appeared in nearly half of its outputs.
More from Models
- Gary Marcus says LLM math skills are like knowing only a car’s engine size — GaryMarcus · 2026-07-22
- OpenAI’s Codex + GPT-5.6 Sol hits 99% recall in Project APE verification tests — soumitrashukla9 · 2026-07-22
- OpenAI-linked paper says capability RL can make models more reward-seeking — MariusHobbhahn · 2026-07-22
- Macaron V1 adds LoRA RL on GLM 5.2 and claims SOTA benchmark gains — Xianbao_QIAN · 2026-07-22
- OpenAI rolls out voice in GPT-Live, but the UI obscures search and reasoning — Graham_dePenros · 2026-07-22
- Moonshot’s Kimi K3 sets a new open-weights ECI record at 156 — scaling01 · 2026-07-22