Frontier models still show greediness and frequency bias, researchers say
m_wulfmeier · x · 2026-07-21
The author says frontier models still show two decision biases: greediness, where the model collapses onto the highest-prior answer, and frequency bias, where salient context outweighs evidence.
They add that some training recipes can reduce these biases over the long term, including RL fine-tuning on self-generated rationales, as shown in their greedy agents paper.
More from Research
- Sakana AI launches Fugu-Cyber and argues benchmark scores are only the start — SakanaAILabs · 2026-07-22
- Meta says SAM 3 and DINOv3 cut 3D volume labeling from a month to 15 minutes — AIatMeta · 2026-07-22
- Project CETI gets a Jeopardy! shout-out with a SETI-style whale clue — begusgasper · 2026-07-22
- Somatic mutations alone may cap human lifespan at 146 to 194 years, study finds — Anen-o-me · 2026-07-22
- Research finds memory compression makes AI agents drop safety rules and hit 59% violations — gerardsans · 2026-07-22
- DriftWorld claims a world model that runs at 30+ FPS and trains on 1–2 GPUs — du_yilun · 2026-07-22