Benchmark Inflation: Revealing LLM Performance Gaps Using Retro-Holdouts
Many suspect that LLMs are 'cheating' on evaluations; to what extent is that accurate?
Jul 24, 2024
Open Source AI is a lie, but it doesn't have to be
Big Tech is attempting to redefine "Open Source" to their advantage; at the very least, we should know about it.
Apr 30, 2024
Detecting Implicit Gaming through Retrospective Evaluation Sets
2-day hackathon project, awarded first place by peer review in the Evaluations Apart Hackathon in Nov '23.
Nov 23, 2023
Into AI Safety Podcast
The podcast was started as a tool for others shifting into the field of AI safety.
Oct 23, 2023
HoPE Against HoPE
Exploring pigeon flocking under predation through multi-agent simulation and evolutionary algorithms.
May 1, 2022
Stretching the Boundary: Shell Finite Elements for Pneumatic Soft Actuators
Leveraging shell finite elements for faster simulation without compromising accuracy.
Apr 8, 2022
Automated Synthesis of Bending Pneumatic Soft Actuators
A seemless workflow for the design and fabrication of pneumatic soft actuators.
Apr 8, 2022
Optimal Design
Reports from course on optimal design, including least squares classification, optimal topology, support vector machines, and an optimization-based Sudoku solver.
Apr 1, 2022
CPPN2Sim
MATLAB library for converting Computational Pattern Producing Networks into lightweight simulations of actuator behavior.
Nov 2, 2021
Flow Visualization
Reports from course on flow visualization: Viscosity Dynamics, Propelled Paint, and Bad Water Rising.
Nov 1, 2021