13 citations · 36 across the 11 of their papers we have counts for
11 papers
Inference economics of language models
Ege Erdil
We develop a theoretical model that addresses the economic trade-off between cost per token versus serial token generation speed when deploying LLMs for inference at scale. Our mod…
GATE: An Integrated Assessment Model for AI Automation
Ege Erdil, Andrei Potlogea, Tamay Besiroglu +6
Assessing the economic impacts of artificial intelligence requires integrating insights from both computer science and economics. We present the Growth and AI Transition Endogenous…
Data movement limits to frontier model training
Ege Erdil, David Schneider-Joseph
We present a theoretical model of distributed training, and use it to analyze how far dense and sparse training runs can be scaled. Under our baseline assumptions, given a three mo…
FrontierMath: A Benchmark for Evaluating Advanced Mathematical Reasoning in AI
Elliot Glazer, Ege Erdil, Tamay Besiroglu +21
We introduce FrontierMath, a benchmark of hundreds of original, exceptionally challenging mathematics problems crafted and vetted by expert mathematicians. The questions cover most…
Estimating Idea Production: A Methodological Survey
Ege Erdil, Tamay Besiroglu, Anson Ho
Accurately modeling the production of new ideas is crucial for innovation theory and endogenous growth models. This paper provides a comprehensive methodological survey of strategi…
Chinchilla Scaling: A replication attempt
Tamay Besiroglu, Ege Erdil, Matthew Barnett +1
Hoffmann et al. (2022) propose three methods for estimating a compute-optimal scaling law. We attempt to replicate their third estimation procedure, which involves fitting a parame…
Algorithmic progress in language models
Anson Ho, Tamay Besiroglu, Ege Erdil +6
We investigate the rate at which algorithms for pre-training language models have improved since the advent of deep learning. Using a dataset of over 200 language model evaluations…
Limits to the Energy Efficiency of CMOS Microprocessors
Anson Ho, Ege Erdil, Tamay Besiroglu
CMOS microprocessors have achieved massive energy efficiency gains but may reach limits soon. This paper presents an approach to estimating the limits on the maximum floating point…
Explosive growth from AI automation: A review of the arguments
Ege Erdil, Tamay Besiroglu
We examine whether substantial AI automation could accelerate global economic growth by about an order of magnitude, akin to the economic growth effects of the Industrial Revolutio…
Power Law Trends in Speedrunning and Machine Learning
Ege Erdil, Jaime Sevilla
We find that improvements in speedrunning world records follow a power law pattern. Using this observation, we answer an outstanding question from previous work: How do we improve…
Algorithmic progress in computer vision
Ege Erdil, Tamay Besiroglu
We investigate algorithmic progress in image classification on ImageNet, perhaps the most well-known test bed for computer vision. We estimate a model, informed by work on neural s…