Cem Anil
21 works · 1,459 citations · h-index 9
Top works
- Training Deep Networks with Synthetic Data: Bridging the Reality Gap by Domain Randomization
- Solving Quantitative Reasoning Problems with Language Models
- Sorting out Lipschitz function approximation
- Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training
- Exploring Length Generalization in Large Language Models
- Many-shot Jailbreaking
- Kronfluence: Influence Functions with Eigenvalue-corrected Kronecker-Factored Approximate Curvature
- Preventing Gradient Attenuation in Lipschitz Constrained Convolutional Networks
- TimbreTron: A WaveNet(CycleGAN(CQT(Audio))) Pipeline for Musical Timbre Transfer
- Connecting the Dots: LLMs can Infer and Verbalize Latent Structure from Disparate Training Data
- Constitutional Classifiers: Defending against Universal Jailbreaks across Thousands of Hours of Red Teaming
- Exploring Length Generalization in Large Language Models
- Path Independent Equilibrium Models Can Better Exploit Test-Time Computation
- Path Independent Equilibrium Models Can Better Exploit Test-Time Computation
- Sabotage Evaluations for Frontier Models
- Modular Pretraining Enables Access Control
- Learning to Elect
- Beyond Data Filtering: Knowledge Localization for Capability Removal in LLMs
- Sparse capsule networks for informative representation learning in digital pathology
- Learning to Give Checkable Answers with Prover-Verifier Games