Released sign-to-text checkpoints combine visual and text-side parameters, so intact-checkpoint comparisons cannot identify what drives transfer. We study this confound in low-resource Taiwan Sign Language (TSL) by separating and recombining the visual and mT5 components of Uni-Sign checkpoints trained on Chinese Sign Language (CSL)-to-Chinese and American Sign Language (ASL)-to-English data, then fine-tuning the combinations identically on TSL. The CSL-derived visual initialization improves content-word accuracy by 15.0 points with its released mT5 parameters, but changes accuracy by only -0.5 points with untouched mT5-base, showing that the measured benefit of visual pretraining depends on the text-side initialization. We assess video dependence by replacing target clips with duration-matched held-out clips at evaluation. Further interventions localize much of the accuracy and video-dependence effects to the untied output projection: replacing an ASL-to-English model’s projection with mT5-base’s improves accuracy by 25.0 points and increases sensitivity to clip replacement. Permutation and timing controls show that the effect depends on pretrained row-to-token correspondence and on the projection being present during fine-tuning. Absolute translation quality remains low, so we interpret these experiments diagnostically. The results show that released checkpoints should be analyzed as structured initializations whose components can interact, rather than treated as indivisible evidence of cross-language transfer.

While the study of diachronic semantic change has advanced alongside recent computational developments, structured lexical resources that reflect semantic evolution remain scarce for many languages, including Ancient Chinese. By systematizing the diachronic transformations within the Chinese Text Project (ctext, a large corpus of Ancient Chinese), we aim to bridge the gap between traditional philological inquiry and contemporary computational linguistics. This study proposes a pipeline that extracts contextualized embeddings from GujiBERT-fan, a language model pre-trained on pre-modern Chinese, and applies dynamic hierarchical clustering to identify distinct senses across historical periods. The pipeline operates at two levels: a global clustering that aggregates data across all periods to capture the full semantic space, and local clustering within each dynasty to reveal period-specific usage patterns. We test the pipeline with a pilot study on the character 手 (shǒu, “hand”) across eight dynastic periods, covering over 185,000 occurrences. The results show that the pipeline can capture the diachronic shift from concrete to abstract senses, demonstrating its potential as a scalable method for mapping semantic evolution in historical languages.

We propose the Linguistics Olympiad Benchmark for Structured Evaluation on Reasoning, or LOBSTER, a linguistically-informed benchmark designed to evaluate large language models (LLMs) on complex linguistic puzzles of the International Linguistics Olympiad (IOL). Unlike prior benchmarks that focus solely on final answer accuracy, our benchmark provides concrete evaluation protocols and rich typological metadata across over 90 low-resource and cross-cultural languages alongside the puzzles. Through systematic evaluations of state-of-the-art models on multilingual abilities, we demonstrate that LLMs struggle with low-resource languages, underscoring the need for such a benchmark. Experiments with various models on our benchmark showed that IOL problems remain a challenging task for reasoning models, though there are ways to enhance the performance—for example, iterative reasoning outperforms single-pass approaches in both final answers and explanations. Our benchmark offers a comprehensive foundation for advancing linguistically grounded, culturally informed, and cognitively plausible reasoning in LLMs.