SimDiff: Advanced Depth Pruning for Efficient LLMs
Discover SimDiff, a novel depth pruning method using similarity and difference metrics to boost large language model efficiency and speed.
Multi-Agent Generative Engine Optimization with Reusable Strategies
Discover MAGEO, a multi-agent framework that optimizes generative engines by reusing strategies for improved visibility and citation fidelity.
CoDA: Enhancing Cross-Domain Knowledge Transfer in LLMs
Discover CoDA, a novel method improving cross-domain knowledge transfer in large language models via CoT-guided domain adaptation and feature distillation.
Evaluating Faithfulness of LLMs in Logical Reasoning
Explore how large language models handle formalization and faithfulness in logical reasoning, revealing key insights on proof validity and model behavior.
Four-Axis Decision Alignment for Enterprise AI Agents
Discover a novel four-axis framework to enhance decision alignment in long-horizon enterprise AI agents for high-stakes, regulated environments.