Despite recent advances in large language models, building dependable and deployable NLP models typically requires abundant, high-quality training data. However, task-specific data is not available for many use cases, and manually curating task-specific data is labor-intensive. Recent work has studied prompt-driven synthetic data generation using large language models, but these generated datasets tend to lack complexity and diversity. To address these limitations, we introduce a method, _DataTune_, to make better use of existing, publicly available datasets to improve automatic dataset generation. DataTune performs _dataset transformation_, enabling the repurposing of publicly available datasets into a format that is directly aligned with the specific requirements of target tasks. On a diverse set of language-based tasks from the BIG-Bench benchmark, we find that finetuning language models via DataTune improves over a few-shot prompting baseline by 49% and improves over existing methods that use synthetic or retrieved training data by 34%. We find that dataset transformation significantly increases the diversity and difficulty of generated data on many tasks. We release a Python package and open-source repository to make this method accessible to the community (URL will be added upon acceptance).

Social media’s ubiquity fosters a space for users to exhibit suicidal thoughts outside of traditional clinical settings. Understanding the build-up of such ideation is critical for the identification of at-risk users and suicide prevention. Suicide ideation is often linked to a history of mental depression. The emotional spectrum of a user’s historical activity on social media can be indicative of their mental state over time. In this work, we focus on identifying suicidal intent in English tweets by augmenting linguistic models with historical context. We propose STATENet, a time-aware transformer based model for preliminary screening of suicidal risk on social media. STATENet outperforms competitive methods, demonstrating the utility of emotional and temporal contextual cues for suicide risk assessment. We discuss the empirical, qualitative, practical, and ethical aspects of STATENet for suicide ideation detection.