Closed-system multi-step LLM reasoning is subject to an information-theoretic bound where mutual information with evidence decreases, preserving accuracy while eroding faithfulness, with EGSR recovering it on SciFact and FEVER.
Educational and Psychological Measurement , volume=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
ConvVitMamba integrates multiscale convolution, transformer encoding, and Mamba-based refinement with PCA to outperform prior CNN, ViT, and Mamba methods in accuracy, size, and speed on four HSI benchmark datasets.
citing papers explorer
-
The Reasoning Trap: An Information-Theoretic Bound on Closed-System Multi-Step LLM Reasoning
Closed-system multi-step LLM reasoning is subject to an information-theoretic bound where mutual information with evidence decreases, preserving accuracy while eroding faithfulness, with EGSR recovering it on SciFact and FEVER.
-
ConvVitMamba: Efficient Multiscale Convolution, Transformer, and Mamba-Based Sequence modelling for Hyperspectral Image Classification
ConvVitMamba integrates multiscale convolution, transformer encoding, and Mamba-based refinement with PCA to outperform prior CNN, ViT, and Mamba methods in accuracy, size, and speed on four HSI benchmark datasets.