This website requires javascript to properly function. Consider activating javascript to get access to all site functionality.
LESSWRONG
Tags
LW
Login
AI Capabilities
•
Applied to
[Paper] AI Sandbagging: Language Models can Strategically Underperform on Evaluations
by
Teun van der Weij
5d
ago
•
Applied to
An Introduction to AI Sandbagging
by
Teun van der Weij
2mo
ago
•
Applied to
Addressing Accusations of Handholding
by
Yeshua God
3mo
ago
•
Applied to
Timelines to Transformative AI: an investigation
by
Zershaaneh Qureshi
3mo
ago
•
Applied to
Self-Play By Analogy
by
the gears to ascension
3mo
ago
•
Applied to
Benchmarking LLM Agents on Kaggle Competitions
by
the gears to ascension
3mo
ago
•
Applied to
Questions I’d Want to Ask an AGI+ to Test Its Understanding of Ethics
by
sweenesm
5mo
ago
•
Applied to
AlphaGeometry: An Olympiad-level AI system for geometry
by
alyssavance
5mo
ago
•
Applied to
AI doing philosophy = AI generating hands?
by
Wei Dai
5mo
ago
•
Applied to
What's the protocol for if a novice has ML ideas that are unlikely to work, but might improve capabilities if they do work?
by
drocta
5mo
ago
•
Applied to
$300 for the best sci-fi prompt: the results
by
RomanS
5mo
ago
•
Applied to
A call for a quantitative report card for AI bioterrorism threat models
by
Juno
7mo
ago
•
Applied to
The Stochastic Parrot Hypothesis is debatable for the last generation of LLMs
by
Quentin FEUILLADE--MONTIXI
7mo
ago
•
Applied to
AI as Super-Demagogue
by
RationalDino
7mo
ago
•
Applied to
[Thought Experiment] Tomorrow's Echo - The future of synthetic companionship.
by
Vimal Naran
8mo
ago
•
Applied to
I Would Have Solved Alignment, But I Was Worried That Would Advance Timelines
by
307th
8mo
ago
•
Applied to
Eleuther releases Llemma: An Open Language Model For Mathematics
by
duck_master
8mo
ago