Search Results for author:
Found 3 papers, 2 papers with code
Multilingual Pretraining and Instruction Tuning Improve Cross-Lingual Knowledge Alignment, But Only Shallowly
In this paper, we propose CLiKA, a systematic framework to assess the cross-lingual knowledge alignment of LLMs in the Performance, Consistency and Conductivity levels, and explored the effect of multilingual pretraining and instruction tuning on the degree of alignment.
Measuring Meaning Composition in the Human Brain with Composition Scores from Large Language Models
Drawing on the key-value memory interpretation of transformer feed-forward network blocks, we introduce the Composition Score, a novel model-based metric designed to quantify the degree of meaning composition during sentence comprehension.
Roles of Scaling and Instruction Tuning in Language Perception: Model vs. Human Attention
Recent large language models (LLMs) have revealed strong abilities to understand natural language.
