We present OTA-BOUN v2.0, the largest Universal Dependencies treebank for historical Turkish, consisting of 1,742 manually verified sentences sampled from late Ottoman texts. The annotation process followed a semi-automatic methodology: initial pre-annotation by the UDPipe 2.0 pipeline was refined through manual annotation of dependency relations, part-of-speech tags, and lemmas. A distinctive feature of OTA-BOUN is its dual-script representation: each sentence is provided both in the original Perso-Arabic script and its Latinized transcription, while tokens include aligned forms in both scripts. This dual-layer design enables research on script conversion, cross-lingual transfer, and historical–modern Turkish comparisons. Through detailed analyses on the aforementioned treebank, this study presents a unique and scalable resource, advancing computational studies of historical Turkish and supporting broader efforts in multilingual and diachronic NLP.

Detecting biases in natural language understanding (NLU) for African American Vernacular English (AAVE) is crucial to developing inclusive natural language processing (NLP) systems. To address dialect-induced performance discrepancies, we introduce AAVENUE (AAVE Natural Language Understanding Evaluation), a benchmark for evaluating large language model (LLM) performance on NLU tasks in AAVE and Standard American English (SAE). AAVENUE builds upon and extends existing benchmarks like VALUE, replacing deterministic syntactic and morphological transformations with a more flexible methodology leveraging LLM-based translation with few-shot prompting, improving performance across our evaluation metrics when translating key tasks from the GLUE and SuperGLUE benchmarks. We compare AAVENUE and VALUE translations using five popular LLMs and a comprehensive set of metrics including fluency, BARTScore, quality, coherence, and understandability. Additionally, we recruit fluent AAVE speakers to validate our translations for authenticity. Our evaluations reveal that LLMs consistently perform better on SAE tasks than AAVE-translated versions, underscoring inherent biases and highlighting the need for more inclusive NLP models.