Sabine Schulte im Walde

This work explores whether LLMs can accurately predict and explain plausible materials for vehicle components such as brake discs or fuel injectors without requiring extensive fine-tuning and substantially outperforms prior work.

Compounds in English and German show only a slight decline in compositionality over decades, challenging long-held assumptions in semantic change research.

Metaphor detection models don't just memorize verbs; they learn transferable contextual cues that allow them to generalize to unseen words.

LLMs will readily twist implausible events into plausible metaphors, revealing a surprising gap in their grasp of real-world constraints compared to humans.

Forget universal rules: distinguishing metaphorical from literal language hinges on construction-specific nuances, not broad distributional patterns.