Update README.md

README.md

@@ -25,11 +25,6 @@ This is a high-quality, executor-validated multi-turn tool-use dataset designed
 | Validation | All trajectories verified against live tool executors |
 | Supported benchmarks | BFCL v4 Multi-Turn, τ²-bench |
 
-## Intended Use
-- Supervised fine-tuning (SFT) of agentic language models for multi-turn tool use
-- Training models to ground tool arguments in verifiable context rather than hallucinating values
-- Evaluation of long-horizon tool-calling capabilities
-
 ## Reference Training Setup
 The dataset was originally used to train a 4B parameter model via pure SFT (no RL):
 - Base model: Qwen3-4B-Instruct