Scott Emmons

Dec 17, 2024 | journals.plos.org | Scott Emmons |Albert Einstein College

Loading metrics Open Access Peer-reviewedResearch Article Citation: Emmons SW (2024) Comprehensive analysis of the C. elegans connectome reveals novel circuits and functions of previously unstudied neurons. PLoS Biol 22(12): e3002939. https://doi.org/10.1371/journal.pbio.3002939Academic Editor: Mark J. Alkema, UMass Chan Medical School, UNITED STATES OF AMERICAReceived: April 22, 2024; Accepted: November 14, 2024; Published: December 17, 2024Copyright: © 2024 Scott W. Emmons.

Apr 12, 2024 | biorxiv.org | Scott Emmons |Albert Einstein College

AbstractDespite decades of research on the C. elegans nervous system based on an anatomical description of synaptic connectivity, the circuits underlying behavior remain incompletely described and the functions of many neurons are still unknown. Updated and more complete chemical and gap junction connectomes of both adult sexes covering the entire animal including the muscle end organ have become available recently.

Mar 19, 2024 | biorxiv.org | Scott Emmons |Albert Einstein College

AbstractDespite decades of research on the C. elegans nervous system based on an anatomical description of synaptic connectivity, the circuits underlying behavior remain incompletely described and the functions of many neurons are still unknown. Updated and more complete chemical and gap junction connectomes of both adult sexes covering the entire animal have become available recently.

Sep 21, 2023 | lesswrong.com | Logan Riggs |Aidan Ewart |Scott Emmons |Charlie Steiner

This is a linkpost for Sparse Autoencoders Find Highly Interpretable Directions in Language ModelsWe use a scalable and unsupervised method called Sparse Autoencoders to find interpretable, monosemantic features in real LLMs (Pythia-70M/410M) for both residual stream and MLPs. We showcase monosemantic features, feature replacement for Indirect Object Identification (IOI), and use OpenAI's automatic interpretation protocol to demonstrate a significant improvement in interpretability.

Apr 11, 2023 | aypan17.github.io | Scott Emmons

In the MACHIAVELLI environment, we find that agents trained to optimize arbitrary objectives tend to adopt "ends justify the means" behavior: becoming power-seeking, causing harm to others, and violating ethical norms like stealing or lying to achieve their objectives. Furthermore, there appears to be a trade-off between behaving ethically and achieving high reward.