manifold-alignment

Tag

Cards List
#manifold-alignment

Hallucinations as Orthogonal Noise: Inference-Time Manifold Alignment via Dynamic Contextual Orthogonalization

arXiv cs.CL · 2026-06-03 Cached

This paper proposes Dynamic Contextual Orthogonalization (DCO), an inference-time method that reduces hallucinations in large language models by aligning attention head outputs with the context manifold, achieving superior faithfulness on benchmarks with Llama-3 models.

0 favorites 0 likes
#manifold-alignment

Self-Distillation as a Performance Recovery Mechanism for LLMs: Counteracting Compression and Catastrophic Forgetting

arXiv cs.CL · 2026-04-20 Cached

This paper introduces Self-Distillation Fine-Tuning (SDFT) as a recovery mechanism for LLMs suffering from performance degradation due to catastrophic forgetting, quantization, and pruning. The authors provide theoretical justification using Centered Kernel Alignment (CKA) to demonstrate that self-distillation aligns the student model's high-dimensional manifold with the teacher's optimal structure, effectively recovering lost capabilities.

0 favorites 0 likes
← Back to home

Submit Feedback