Picture for Evan Hubinger

Evan Hubinger

Discovering Language Model Behaviors with Model-Written Evaluations

Dec 19, 2022
Ethan Perez, Sam Ringer, Kamilė Lukošiūtė, Karina Nguyen, Edwin Chen, Scott Heiner, Craig Pettit, Catherine Olsson, Sandipan Kundu, Saurav Kadavath, Andy Jones, Anna Chen, Ben Mann, Brian Israel, Bryan Seethor, Cameron McKinnon, Christopher Olah, Da Yan, Daniela Amodei, Dario Amodei, Dawn Drain, Dustin Li, Eli Tran-Johnson, Guro Khundadze, Jackson Kernion, James Landis, Jamie Kerr, Jared Mueller, Jeeyoon Hyun, Joshua Landau, Kamal Ndousse, Landon Goldberg, Liane Lovitt, Martin Lucas, Michael Sellitto, Miranda Zhang, Neerav Kingsland, Nelson Elhage, Nicholas Joseph, Noemí Mercado, Nova DasSarma, Oliver Rausch, Robin Larson, Sam McCandlish, Scott Johnston, Shauna Kravec, Sheer El Showk, Tamera Lanham, Timothy Telleen-Lawton, Tom Brown, Tom Henighan, Tristan Hume, Yuntao Bai, Zac Hatfield-Dodds, Jack Clark, Samuel R. Bowman, Amanda Askell, Roger Grosse, Danny Hernandez, Deep Ganguli, Evan Hubinger, Nicholas Schiefer, Jared Kaplan

Figure 1 for Discovering Language Model Behaviors with Model-Written Evaluations

Figure 2 for Discovering Language Model Behaviors with Model-Written Evaluations

Figure 3 for Discovering Language Model Behaviors with Model-Written Evaluations

Figure 4 for Discovering Language Model Behaviors with Model-Written Evaluations

* for associated data visualizations, see https://www.evals.anthropic.com/model-written/ for full datasets, see https://github.com/anthropics/evals

Access Paper or Ask Questions

Engineering Monosemanticity in Toy Models

Nov 16, 2022
Adam S. Jermyn, Nicholas Schiefer, Evan Hubinger

Figure 1 for Engineering Monosemanticity in Toy Models

Figure 2 for Engineering Monosemanticity in Toy Models

Figure 3 for Engineering Monosemanticity in Toy Models

Figure 4 for Engineering Monosemanticity in Toy Models

* 31 pages, 26 figures

Access Paper or Ask Questions

An overview of 11 proposals for building safe advanced AI

Dec 04, 2020
Evan Hubinger

Figure 1 for An overview of 11 proposals for building safe advanced AI

Figure 2 for An overview of 11 proposals for building safe advanced AI

Figure 3 for An overview of 11 proposals for building safe advanced AI

Figure 4 for An overview of 11 proposals for building safe advanced AI

Access Paper or Ask Questions