But now we also need new tools and techniques to audit intent. The last thing we want is models turning on us ‘for our own good’.
show comments
drivebyhooting
In terms of safety and alignment, how is this different from looping or increasing the depth?
Fe2O3
This seems really cool, and it actually increases intelligence than just speeding up inference. How does it convert latent reasoning into human readable text?
show comments
jedbrooke
maybe the font experts can check, but I’m pretty sure they’re using comic sans for the text in the diagrams :)
show comments
44za12
Where are the benchmarks? Does it just speed up inference and reduce token usage or does it seem to improve accuracy on certain tasks as well?
The code is available at: https://github.com/togethercomputer/HybridAgent
LLM Models and data: https://huggingface.co/collections/togethercomputer/hybridag...
Evaluations: https://arxiv.org/html/2610.18861v1#S5
Oh cool, it's the Most Forbidden Technique
How does this differ from Astra's looped transformer? That architecture also reasons in the latent space
https://x.com/Jiaru_Zou/status/2110400256647340476
https://github.com/togethercomputer/HybridAgent
This is the way.
But now we also need new tools and techniques to audit intent. The last thing we want is models turning on us ‘for our own good’.
In terms of safety and alignment, how is this different from looping or increasing the depth?
This seems really cool, and it actually increases intelligence than just speeding up inference. How does it convert latent reasoning into human readable text?
maybe the font experts can check, but I’m pretty sure they’re using comic sans for the text in the diagrams :)
Where are the benchmarks? Does it just speed up inference and reduce token usage or does it seem to improve accuracy on certain tasks as well?