Making Large Language Models Speak Tulu: Structured Prompting for an Extremely Low-Resource Language
This addresses the challenge of making AI accessible for speakers of low-resource languages, though it is incremental as it builds on existing prompting methods rather than introducing a fundamentally new approach.
The researchers tackled the problem of enabling large language models to converse in Tulu, an extremely low-resource language with minimal digital presence, by using structured prompting techniques without fine-tuning. Their approach reduced vocabulary contamination from 80% to 5% and achieved 85% grammatical accuracy across three LLMs, with negative constraints providing consistent improvements of 12-18 percentage points.
Can large language models converse in languages virtually absent from their training data? We investigate this question through a case study on Tulu, a Dravidian language with over 2 million speakers but minimal digital presence. Rather than fine-tuning an LLM, we examine whether structured prompts alone can elicit basic conversational ability under controlled prompting. We systematically tackle various challenges posed by absence of training data for Tulu by combining explicit grammar documentation, negative constraints to suppress high-probability tokens from related languages, romanization standardization, and quality-controlled synthetic data generation via self-play. Evaluated on a manually curated held-out set across three LLMs (Gemini 2.0 Flash, GPT-4o, Llama 3.1 70B) and validated by native speakers, our approach reduces vocabulary contamination from 80% to 5% while achieving 85% grammatical accuracy. Cross-model analysis reveals that negative constraints provide consistent improvements (12--18 percentage points), while grammar documentation effects vary by model architecture (8--22 points).