SAGE: Scalable Ground Truth Evaluations for Large Sparse Autoencoders

SAGE: Scalable Ground Truth Evaluations for Large Sparse Autoencoders

9 October 2024

Constantin Venhoff

Anisoara Calinescu

Christian Schroeder de Witt

ArXiv (abs)PDF HTML

Papers citing "SAGE: Scalable Ground Truth Evaluations for Large Sparse Autoencoders"

13 / 13 papers shown

Title
Measuring Progress in Dictionary Learning for Language Model Interpretability with Board Game Models Adam Karvonen Benjamin Wright Can Rager Rico Angell Jannik Brinkmann Logan Smith C. M. Verdun David Bau Samuel Marks 70 31 0 31 Jul 2024
Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models Samuel Marks Can Rager Eric J. Michaud Yonatan Belinkov David Bau Aaron Mueller 133 158 0 28 Mar 2024
Attribution Patching Outperforms Automated Circuit Discovery Aaquib Syed Can Rager Arthur Conmy 135 67 0 16 Oct 2023
Frontier AI Regulation: Managing Emerging Risks to Public Safety Markus Anderljung Joslyn Barnhart Anton Korinek Jade Leung Cullen O'Keefe ... Jonas Schuett Yonadav Shavit Divya Siddarth Robert F. Trager Kevin J. Wolf SILM 88 125 0 06 Jul 2023
Towards Automated Circuit Discovery for Mechanistic Interpretability Arthur Conmy Augustine N. Mavor-Parker Aengus Lynch Stefan Heimersheim Adrià Garriga-Alonso 66 319 0 28 Apr 2023
Interpreting Neural Networks through the Polytope Lens Sid Black Lee D. Sharkey Léo Grinsztajn Eric Winsor Daniel A. Braun ... Kip Parker Carlos Ramón Guevara Beren Millidge Gabriel Alfour Connor Leahy FAtt MILM 70 26 0 22 Nov 2022
Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small Kevin Wang Alexandre Variengien Arthur Conmy Buck Shlegeris Jacob Steinhardt 310 563 0 01 Nov 2022
In-context Learning and Induction Heads Catherine Olsson Nelson Elhage Neel Nanda Nicholas Joseph Nova Dassarma ... Tom B. Brown Jack Clark Jared Kaplan Sam McCandlish C. Olah 319 528 0 24 Sep 2022
X-Risk Analysis for AI Research Dan Hendrycks Mantas Mazeika 69 71 0 13 Jun 2022
Language Models are Few-Shot Learners Tom B. Brown Benjamin Mann Nick Ryder Melanie Subbiah Jared Kaplan ... Christopher Berner Sam McCandlish Alec Radford Ilya Sutskever Dario Amodei BDL 880 42,463 0 28 May 2020
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer Colin Raffel Noam M. Shazeer Adam Roberts Katherine Lee Sharan Narang Michael Matena Yanqi Zhou Wei Li Peter J. Liu AIMat 485 20,342 0 23 Oct 2019
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding Jacob Devlin Ming-Wei Chang Kenton Lee Kristina Toutanova VLM SSL SSeg 1.8K 95,229 0 11 Oct 2018
Axiomatic Attribution for Deep Networks Mukund Sundararajan Ankur Taly Qiqi Yan OOD FAtt 193 6,024 0 04 Mar 2017