Articles by bearseascape
12

Transformers Are Inherently Succinct (arxiv.org)

3

Slople – Can you tell real ML papers from AI-generated ones? (ml5885.github.io)

1

Benchmarking Culture (argmin.net)

4

Why one small American town won't stop stoning its residents to death (archiveofourown.org)

1

The most complex model we understand [video] (youtube.com)

1

Weird Generalization and Inductive Backdoors: New Ways to Corrupt LLMs (arxiv.org)

3

MooseAgent: A LLM Based Multi-Agent Framework for Automating Moose Simulation (arxiv.org)

2

Automated Researchers Can Subtly Sandbag (anthropic.com)

1

Auditing Language Models for Hidden Objectives (anthropic.com)

1

Policy for LLM Writing on LessWrong (lesswrong.com)

1

Towards Understanding Distilled Reasoning Models: A Representational Approach (arxiv.org)

1

(Mis)Fitting: A Survey of Scaling Laws (arxiv.org)

1

Resurrecting saturated LLM benchmarks with adversarial encoding (arxiv.org)

1

Deep Double Descent: Where Bigger Models and More Data Hurt (openai.com)

18

Value-Based Deep RL Scales Predictably (arxiv.org)