Picture for Rylan Schaeffer

Rylan Schaeffer

Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive?

Add code
Jun 06, 2024
Viaarxiv icon

Is Model Collapse Inevitable? Breaking the Curse of Recursion by Accumulating Real and Synthetic Data

Apr 01, 2024
Viaarxiv icon

Bridging Associative Memory and Probabilistic Modeling

Feb 15, 2024
Viaarxiv icon

Investigating Data Contamination for Pre-training Language Models

Add code
Jan 11, 2024
Viaarxiv icon

Incidental Polysemanticity

Dec 05, 2023
Viaarxiv icon

Self-Supervised Learning of Representations for Space Generates Multi-Modular Grid Cells

Nov 04, 2023
Viaarxiv icon

Pretraining on the Test Set Is All You Need

Sep 13, 2023
Viaarxiv icon

Deceptive Alignment Monitoring

Jul 26, 2023
Viaarxiv icon

Invalid Logic, Equivalent Gains: The Bizarreness of Reasoning in Language Model Prompting

Jul 23, 2023
Figure 1 for Invalid Logic, Equivalent Gains: The Bizarreness of Reasoning in Language Model Prompting
Figure 2 for Invalid Logic, Equivalent Gains: The Bizarreness of Reasoning in Language Model Prompting
Figure 3 for Invalid Logic, Equivalent Gains: The Bizarreness of Reasoning in Language Model Prompting
Figure 4 for Invalid Logic, Equivalent Gains: The Bizarreness of Reasoning in Language Model Prompting
Viaarxiv icon

FACADE: A Framework for Adversarial Circuit Anomaly Detection and Evaluation

Jul 20, 2023
Figure 1 for FACADE: A Framework for Adversarial Circuit Anomaly Detection and Evaluation
Viaarxiv icon