AI evaluation | Tech
Tech news, context and sources.
Latest articles
PsAIch research finds AI anxiety narratives change with question framing
High scores on human psychological scales should not be read as evidence that a model experiences distress or consciousness.
Top model passes 10.8% of tasks in DayJob Tax agent benchmark
Recognizing unstated requirements and conflicting documents remains a major professional-judgment challenge.
Artificial Analysis v5 moves 60% of its index to private evaluations
Changed test composition and not-yet-released model weights complicate headline comparisons.
You’ve reached the end.