9 citations · 15 across the 11 of their papers we have counts for
12 papers
Math for AI safety: an invitation for mathematicians
Lionel Levine
Artificial intelligence threatens to outrun human understanding and control. New mathematics is needed to design AI that is legible, steerable, and cooperative with humanity. I org…
Mixing Times of Glauber Dynamics on Masked Language Models
Suvadip Sana, Sami Wolf, Neer Mehta +4
Masked language models (MLMs) define local conditional distributions over tokens but do not, in general, correspond to any consistent joint distribution over sequences. This raises…
On Safety Risks in Experience-Driven Self-Evolving Agents
Weixiang Zhao, Yichen Zhang, Yingshuo Wang +8
Experience-driven self-evolution has emerged as a promising paradigm for improving the autonomy of large language model agents, yet its reliance on self-curated experience introduc…
MI-to-Mid Distilled Compression (M2M-DC): An Hybrid-Information-Guided-Block Pruning with Progressive Inner Slicing Approach to Model Compression
Lionel Levine, Haniyeh Ehsani Oskouie, Sajjad Ghiasvand +1
We introduce MI-to-Mid Distilled Compression (M2M-DC), a two-scale, shape-safe compression framework that interleaves information-guided block pruning with progressive inner slicin…
A Definition of AGI
Dan Hendrycks, Dawn Song, Christian Szegedy +30
The lack of a concrete definition for Artificial General Intelligence (AGI) obscures the gap between today's specialized AI and human-level cognition. This paper introduces a quant…
EigenBench: A Comparative Behavioral Measure of Value Alignment
Jonathn Chang, Leonhard Piff, Suvadip Sana +2
Aligning AI with human values is a pressing unsolved problem. To address the lack of quantitative metrics for value alignment, we propose EigenBench: a black-box method for compara…
Exploring the Impact of Dataset Statistical Effect Size on Model Performance and Data Sample Size Sufficiency
Arya Hatamian, Lionel Levine, Haniyeh Ehsani Oskouie +1
Having a sufficient quantity of quality data is a critical enabler of training effective machine learning models. Being able to effectively determine the adequacy of a dataset prio…
How to quantify the coherence of a set of beliefs
Rowan Hess, Lionel Levine
Given conflicting probability estimates for a set of events, how can we quantify how much they conflict? How can we find a single probability distribution that best encapsulates th…
Locally Markov walks on finite graphs
Robin Kaiser, Lionel Levine, Ecaterina Sava-Huss
Locally Markov walks are natural generalizations of classical Markov chains, where instead of a particle moving independently of the past, it decides where to move next depending o…
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…
Exploring Cross-model Neuronal Correlations in the Context of Predicting Model Performance and Generalizability
Haniyeh Ehsani Oskouie, Sajjad Ghiasvand, Lionel Levine +1
As Artificial Intelligence (AI) models are increasingly integrated into critical systems, the need for a robust framework to establish the trustworthiness of AI is increasingly par…
Do language models plan ahead for future tokens?
Wilson Wu, John X. Morris, Lionel Levine
Do transformers "think ahead" during inference at a given position? It is known transformers prepare information in the hidden states of the forward pass at time step that is t…