curius graph

all pages

showing 80601-80650 of 203113 pages (sorted by popularity)
Constellations are Younger than Continents — LessWrong
Deceptive Alignment is <1% Likely by Default — LessWrong
The Flask Mega-Tutorial, Part II: Templates - miguelgrinberg.com
The Flask Mega-Tutorial, Part XVII: Deployment on Linux - miguelgrinberg.com
The Flask Mega-Tutorial, Part XIV: Ajax - miguelgrinberg.com
My take on Jacob Cannell’s take on AGI safety — LessWrong
AGI ruin scenarios are likely (and disjunctive) — LessWrong
AI Risk and the US Presidential Candidates — LessWrong
Arbital
The sin of updating when you can change whether you exist — LessWrong
A transparency and interpretability tech tree — AI Alignment Forum
Transposed Convolutions explained with… MS Excel! | by Thom Lane | Apache MXNet | Medium
Copy of [0.5] GANs & VAEs (exercises).ipynb - Colaboratory
Clothing For Men — LessWrong
On Anthropic's Sleeper Agents Paper - by Zvi Mowshowitz
Gender Imbalances Are Mostly Not Due To Offensive Attitudes | Slate Star Codex
[2310.15421] FANToM: A Benchmark for Stress-testing Machine Theory of Mind in Interactions
in praise of uselessness - by Ava - bookbear express
Coup probes: Catching catastrophes with probes trained off-policy — LessWrong
Rescuing the utility function - Arbital
The Hidden Complexity of Wishes — LessWrong
Direct Preference Optimization (DPO) | by João Lages | Medium
[2312.08358] Distributional Preference Learning: Understanding and Accounting for Hidden Context in RLHF
Chinese Coercion in the South China Sea: Resolve and Costs | Belfer Center for Science and International Affairs
Latent Adversarial Training — LessWrong
Israel-Hamas war: The US needs to update its old thinking - Vox
Israel's two wars - by Matthew Yglesias - Slow Boring
However Difficult, The United States Should Still Pursue Israeli-Palestinian Peace - War on the Rocks
Rush by west to back Israel erodes developing countries’ support for Ukraine
Sparsify: A mechanistic interpretability research agenda — AI Alignment Forum
Mechanistic anomaly detection and ELK — LessWrong
Primer on Safety Standards and Regulations for Industrial-Scale AI Development – BlueDot Impact
Learning Diverse Skills via Maximum Entropy Deep Reinforcement Learning – The Berkeley Artificial Intelligence Research Blog
[Interim research report] Activation plateaus & sensitive directions in GPT2 — LessWrong
1. The CAST Strategy — LessWrong
Multi-Component Learning and S-Curves — LessWrong
KL Divergence: Forward vs Reverse? - Agustinus Kristiadi
Core Pathways of Aging - LessWrong
On the Algorithmic Bias of Aligning Large Language Models with RLHF: Preference Collapse and Matching Regularization
Recovering the Pre-Fine-Tuning Weights of Generative Models
GPT-2's positional embedding matrix is a helix — LessWrong
Unexpected Benefits of Self-Modeling in Neural Systems
DSLT 2. Why Neural Networks obey Occam's Razor — LessWrong
Miki AOYAGI | Professor (Associate) | Nihon University, Tokyo | Nichidai | Department of Mathematics | Research profile
DSLT 3. Neural Networks are Singular — LessWrong
Decision theory and dynamic inconsistency — LessWrong
(Approximately) Deterministic Natural Latents — LessWrong
Formal verification, heuristic explanations and surprise accounting — Alignment Research Center
Adding Integers in Logarithmic Time | Tim Mastny
Research update: Towards a Law of Iterated Expectations for Heuristic Estimators — LessWrong