Generalized Wick Decompositions von MacLeod, Chris, Nitishinskaya, Evgenia, Shlegeris, Buck

Volltext bestellen

Games for AI Control: Models of Safety Evaluations of AI Deployment Protocols von Griffin, Charlie, Thomson, Louis, Shlegeris, Buck, Abate, Alessandro

Volltext bestellen

AI Control: Improving Safety Despite Intentional Subversion von Greenblatt, Ryan, Shlegeris, Buck, Sachan, Kshitij, Roger, Fabien

Volltext bestellen

Benchmarks for Detecting Measurement Tampering von Roger, Fabien, Greenblatt, Ryan, Nadeau, Max, Shlegeris, Buck, Thomas, Nate

Volltext bestellen

Language models are better than humans at next-token prediction von Shlegeris, Buck, Roger, Fabien, Chan, Lawrence, McLean, Euan

Volltext bestellen

Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small von Wang, Kevin, Variengien, Alexandre, Conmy, Arthur, Shlegeris, Buck, Steinhardt, Jacob

Volltext bestellen

Generalized Wick Decompositions von MacLeod, Chris, Nitishinskaya, Evgenia, Buck Shlegeris

Volltext

Adaptive Deployment of Untrusted LLMs Reduces Distributed Threats von Wen, Jiaxin, Hebbar, Vivek, Larson, Caleb, Bhatt, Aryan, Radhakrishnan, Ansh, Sharma, Mrinank, Sleight, Henry, Feng, Shi, He, He, Perez, Ethan, Shlegeris, Buck, Khan, Akbir

Volltext bestellen

Polysemanticity and Capacity in Neural Networks von Scherlis, Adam, Sachan, Kshitij, Jermyn, Adam S, Benton, Joe, Shlegeris, Buck

Volltext bestellen

Games for AI Control: Models of Safety Evaluations of AI Deployment Protocols von Griffin, Charlie, Thomson, Louis, Buck Shlegeris, Abate, Alessandro

Volltext

AI Control: Improving Safety Despite Intentional Subversion von Greenblatt, Ryan, Buck Shlegeris, Sachan, Kshitij, Fabien, Roger

Volltext

Language models are better than humans at next-token prediction von Buck Shlegeris, Fabien, Roger, Chan, Lawrence, McLean, Euan

Volltext

Supervising strong learners by amplifying weak experts von Christiano, Paul, Shlegeris, Buck, Amodei, Dario

Volltext bestellen

Towards evaluations-based safety cases for AI scheming von Balesni, Mikita, Hobbhahn, Marius, Lindner, David, Meinke, Alexander, Korbak, Tomek, Clymer, Joshua, Shlegeris, Buck, Scheurer, Jérémy, Stix, Charlotte, Shah, Rusheb, Goldowsky-Dill, Nicholas, Braun, Dan, Chughtai, Bilal, Evans, Owain, Kokotajlo, Daniel, Bushnaq, Lucius

Volltext bestellen

Sabotage Evaluations for Frontier Models von Benton, Joe, Wagner, Misha, Christiansen, Eric, Anil, Cem, Perez, Ethan, Srivastav, Jai, Durmus, Esin, Ganguli, Deep, Kravec, Shauna, Shlegeris, Buck, Kaplan, Jared, Karnofsky, Holden, Hubinger, Evan, Grosse, Roger, Bowman, Samuel R, Duvenaud, David

Volltext bestellen

Sycophancy to Subterfuge: Investigating Reward-Tampering in Large Language Models von Denison, Carson, MacDiarmid, Monte, Barez, Fazl, Duvenaud, David, Kravec, Shauna, Marks, Samuel, Schiefer, Nicholas, Soklaski, Ryan, Tamkin, Alex, Kaplan, Jared, Shlegeris, Buck, Bowman, Samuel R, Perez, Ethan, Hubinger, Evan

Volltext bestellen

Benchmarks for Detecting Measurement Tampering von Fabien, Roger, Greenblatt, Ryan, Nadeau, Max, Buck Shlegeris, Thomas, Nate

Volltext

Adversarial Training for High-Stakes Reliability von Ziegler, Daniel M, Nix, Seraphina, Chan, Lawrence, Bauman, Tim, Schmidt-Nielsen, Peter, Lin, Tao, Scherlis, Adam, Nabeshima, Noa, Weinstein-Raun, Ben, de Haas, Daniel, Shlegeris, Buck, Thomas, Nate

Volltext bestellen

Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small von Wang, Kevin, Variengien, Alexandre, Conmy, Arthur, Buck Shlegeris, Steinhardt, Jacob

Volltext

Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training von Hubinger, Evan, Denison, Carson, Mu, Jesse, Lambert, Mike, Tong, Meg, MacDiarmid, Monte, Lanham, Tamera, Ziegler, Daniel M, Maxwell, Tim, Cheng, Newton, Jermyn, Adam, Askell, Amanda, Radhakrishnan, Ansh, Anil, Cem, Duvenaud, David, Ganguli, Deep, Barez, Fazl, Clark, Jack, Ndousse, Kamal, Sachan, Kshitij, Sellitto, Michael, Sharma, Mrinank, DasSarma, Nova, Grosse, Roger, Kravec, Shauna, Bai, Yuntao, Witten, Zachary, Favaro, Marina, Brauner, Jan, Karnofsky, Holden, Christiano, Paul, Bowman, Samuel R, Graham, Logan, Kaplan, Jared, Mindermann, Sören, Greenblatt, Ryan, Shlegeris, Buck, Schiefer, Nicholas, Perez, Ethan

Volltext bestellen