6 papers
Are Single-Token Sparse Autoencoder Features Causally Necessary? Layer-Depth and SAE-Family Effects
Seonglae Cho, Zekun Wu, Kleyton Da Costa +3
Sparse autoencoder (SAE) features are used to interpret and steer large language models, yet nobody has tested whether a feature's causal role is stable across SAE families. Single…
Automata from Agent Traces: Failure and Next-Step Prediction
Seonglae Cho, Franklin Cardenoso Fernandez, Umar Mohammed +4
LLM-based agents execute multi-step tasks, but their behavioral structure remains opaque: long unstructured traces resist the safety auditing and runtime monitoring that deployment…
When Should Graph Attention Be Sparse? Learning a Per-Edge Tsallis Index
Kleyton da Costa, Bernardo Modenesi
Graph attention normalizes neighborhood scores with softmax, the maximum-entropy choice under Shannon statistics. But homophilic and heterophilic graphs want different attention sh…
Perspectives on Tsallis Statistics for Artificial Intelligence
Kleyton da Costa, Bernardo Modenesi
Tsallis statistics generalizes Boltzmann-Gibbs statistical mechanics through a single real parameter that controls the weight assigned to rare and frequent events. Originally p…
GraphNetz: Statistical Benchmarking of Graph Neural Networks with Paired Tests and Rank Aggregation
Kleyton da Costa, Bernardo Modenesi
Graph Neural Networks (GNNs) benchmarks often report single point estimates, even when performance differences are small relative to variation across random seeds, train/test split…
The Confidence Manifold: Geometric Structure of Correctness Representations in Language Models
Seonglae Cho, Zekun Wu, Kleyton Da Costa +1
When a language model asserts that "the capital of Australia is Sydney," does it know this is wrong? Models assert misconceptions with the same fluency as facts, so the question ca…