Tracey Mills

PhD student @MITCoCoSci

[Ionatan Kuperwajs @Ikuperwajs]

If you’ll be at #CogSci2025, consider attending our workshop on meta-reasoning! Organized with Lio Wong, @traceymills_, @katie_m_collins, and @cocosci_lab, we’ll be discussing problem selection through various lenses represented by our speakers 🤔💭 https://t.co/KVMpm1BP8w

[Tracey Mills @traceymills_]

@gabe_grand @OpenAI Lol only slightly ominous 😶

[Tracey Mills @traceymills_]

8/ If we generate instances of a certain kind by drawing from relevant locations in representational space… why these locations? We suggest they are generally useful in our experiences with the category in question -- more in the paper!

[Tracey Mills @traceymills_]

Very excited to have this out w/ @phillipsjs! 🧵⬇️

[Cognition @CognitionJourn]

Think of a zoo animal. Why that one? @traceymills_ & @phillipsjs explain why some category members come to mind over others, based on their locations in category-specific feature space

https://t.co/WXUKz6Djlj

[Tracey Mills @traceymills_]

7/ This holds during decision making as well: people consider options which appear to be drawn from these relevant feature space locations, but select options with different, more context-appropriate features