Tootfinder

RADAR: Mechanistic Pathways for Detecting Data Contamination in LLM Evaluation
Ashish Kattamuri, Harshwardhan Fartale, Arpita Vats, Rahul Raja, Ishita Prasad
https://arxiv.org/abs/2510.08931

RADAR: Mechanistic Pathways for Detecting Data Contamination in LLM Evaluation
Data contamination poses a significant challenge to reliable LLM evaluation, where models may achieve high performance by memorizing training data rather than demonstrating genuine reasoning capabilities. We introduce RADAR (Recall vs. Reasoning Detection through Activation Representation), a novel framework that leverages mechanistic interpretability to detect contamination by distinguishing recall-based from reasoning-based model responses. RADAR extracts 37 features spanning surface-level co…

Now we’re talking 😍

SymTFT Entanglement and Holographic (Non)-Factorization
Ethan Torres, Xingyang Yu
https://arxiv.org/abs/2510.06319 https://arxiv.org/pdf/2510.06319

SymTFT Entanglement and Holographic (Non)-Factorization
Given two otherwise decoupled $D$-dimensional CFTs which possess a common (finite) symmetry subcategory, one can consider entangled boundary states of their $(D+1)$-dimensional SymTFTs. This roughly corresponds to performing a gauging of the tensor product of two CFTs, and we call this phenomena ``SymTFT entanglement" (or ``S-entanglement" for short). In the case when these CFTs have semiclassical holographic duals, the S-entanglement relates the bulk gauge charges between two otherwise disconn…