Lijie Hu

SFAD achieves a remarkable 2.48x speedup in inference while enhancing contextual faithfulness, addressing one of the most pressing challenges in large language models.

Standard attention mechanisms inevitably cause intertask interference in in-context continual learning, leading to systematic bias and performance degradation in long prompts.

Contrastive learning's entanglement problem may be solved: BayesNCL disentangles representations by selectively filtering task-irrelevant features, leading to a 142% boost in semantic consistency.

Applying representation interventions adaptively based on input characteristics dramatically improves alignment without sacrificing general capabilities, a feat previously unmet by uniform intervention methods.

The best LLM to answer a question isn't always the best LLM to *teach* the answer, and matching the "difficulty" of the explanation to the student's current abilities yields better learning.

By enforcing graph isomorphism across counterfactual inputs, UGID reveals that debiasing LLMs can be achieved by directly manipulating internal representations and attention mechanisms.

LLM watermarks can now survive fine-tuning, quantization, and distillation thanks to a new method that embeds them in a stable functional subspace.