Demystifying Entropy-based Selection for Chain-of-Thought Compression in Large Reasoning Models
Paper • 2607.28707 • Published
AI & ML interests
Trustworthy & explainable AI. Neuro-explicit and neurosymbolic methods. Statistical ML.
Papers
RogueAI: A Reverse Turing Test for Detecting Licensed AI Deception in Dialogue
Blending adversarial training and representation-conditional purification via aggregation improves adversarial robustness
saracandu
authored a
paper 2 months ago
emaballarin
updated a
collection 3 months ago
saracandu
updated a
collection 3 months ago
saracandu
authored 2
papers 3 months ago
saracandu
authored 3
papers 4 months ago
Bridging Logic and Learning: Decoding Temporal Logic Embeddings via Transformers
Paper • 2507.07808 • Published
A Dialectic Pipeline for Improving LLM Robustness
Paper • 2601.20659 • Published
Distilling Formal Logic into Neural Spaces: A Kernel Alignment Approach for Signal Temporal Logic
Paper • 2603.05198 • Published
emaballarin
authored 2
papers 4 months ago