| name | tool-use-training-multilingual |
| title | Teaching a Language Model to Speak the Language of Tools |
| version | 0.0.2 |
| engine | skillxiv-v0.0.2-claude-opus-4.6 |
| license | MIT |
| url | https://arxiv.org/abs/2506.23394 |
| keywords | ["Tool Use","Function Calling","Multilingual Models","Instruction Tuning","Language Model Adaptation"] |
| description | Enable non-English language models to reliably generate function calls and tool invocations through continued training on bilingual function-calling datasets. Achieves 28% improvement in parsing accuracy while preserving language understanding. |
Tool-Use Training: Enabling Function Calls in Non-English Language Models
Most language models with reliable tool-use capabilities are English-dominant, leaving multilingual and low-resource language models unable to make clean function calls. When you ask a Bulgarian or Japanese model to use a tool, it often responds with verbose natural language explanations instead of parsable structured output. This forces developers to use English models regardless of their language requirements, limiting tool-augmented AI to English-speaking users.
The core problem is that tool-use training is typically English-centric. Models learn to generate function calls by being fine-tuned on English function-calling examples, which creates a language-specific skill. Teaching a model to reliably generate tool calls in another language requires explicitly training it on function-calling examples in that language, paired with English to show the relationship between natural language requests and structured outputs.
Core Concept
Tool-use training works by treating function calls as a special output format that the model must learn to produce reliably. The key insight is:
- Function calls are a structured language with precise syntax requirements
- Non-English models can learn this structure through exposure to bilingual examples
- Clean, parsable output (proper JSON, proper argument order) matters more than natural language explanations
- Continued training on domain-specific function-calling data, even with relatively small datasets, significantly improves accuracy
The approach uses bilingual training data where the same functional intent appears in both English and the target language, showing the model that "call this function" should produce the same structured output regardless of input language.
Architecture Overview
Tool-use training leverages the model's existing architecture without modification. The training pipeline consists of:
- Bilingual Dataset: Function-calling examples with English and target-language prompts paired with identical function call outputs
- Continued Training Phase: Fine-tune the base model on function-calling data with standard language modeling loss
- Instruction Tuning Integration: Mix function-calling examples with general instruction examples to preserve general capabilities
- Parsing Validation: Post-generation filtering that enforces valid JSON/structured output format
Implementation
Step 1: Create bilingual function-calling dataset
Build training data where the same function call appears in both English and the target language. Start with existing MCP (Model Context Protocol) specifications or tool definitions.
