Tried feeding Golden Gate Claude emojis.
@petitegeek
-
RLHF Hard-Coding: How AI Models Learn Topic Avoidance
By
–
Fascinating! I did find a prompt that didn't reference it at all, does it mean that the topic is too far away to Golden Gate Bridge, or that the question/answer has been hard-coded or strongly RLHF'd?
-
MTurk Survey Results on AI: Main Public Responses
By
–
They asked >1000 people on MTurk, and here were the main responses.
-
Embodied AI Tutors and Virtual Companions in Fiction
By
–
+1 to Klara and the Sun, it explores embodied AI tutors and friends for kids (learned about it thanks to a citation alert linking my research to the novel…) Joi from Blade Runner 2049 was a holographic virtual companion whose existence was tied to a memory stick.
-
GPT-4V Emotion Recognition Benchmarks and Performance Analysis
By
–
Some benchmarks on GPT-4v and emotion recognition in context http://
arxiv.org/abs/2405.08992 -
Next Generation Voice Interfaces: Adapting Technology to Human Expression
By
–
Maybe the next generation will go back to voice 🙂 I really hope we can preserve human ways of expressivity by making the devices adapt to us, instead of us adapt to the devices… Sounds like the latter has already taken this generation!
-
Voice AI for daily tasks: email, music, and future event discovery
By
–
For voice, I often use it to type my emails + slack messages because it's much faster than typing. I use it to play/change music at home. In the future, I really hope it can summarize the kid-friendly events on the weekend so I don't have to scroll instagram. Not there yet tho 🙂
-
OpenAI UI Performance: Latency and Error Handling Analysis
By
–
But is that because the latency and error rate+handling has been bad up until now? I am not a fan of the OpenAI personality, but the UI interaction seems fantastic and very smooth.
