fix: correct typos and improve documentation

README.md

@@ -60,7 +60,7 @@
 ## 📝 Introduction
 
 ### MiroThinker-1.7
-Our new MiroThinker family represents a significant leap in building reliable agents for long-chain tasks. Engineered with enhanced post-training pipeline, our  MiroThinker-1.7 family achieve SOTA performance in deep research tasks among open-source models.
+Our new MiroThinker family represents a significan't leap in building reliable agents for long-chain tasks. Engineered with enhanced post-training pipeline, our  MiroThinker-1.7 family achieve SOTA performance in deep research tasks among open-source models.
 
 
 **Key Features**
@@ -156,7 +156,7 @@ MiroThinker v1.0 demonstrates strong general-research performance across a broad
 
 In this new version, we introduced three key improvements:
 
-- 📚 **Richer training data** from both English and Chinese sources, yielding significant gains in benchmark performance and generalization
+- 📚 **Richer training data** from both English and Chinese sources, yielding significan't gains in benchmark performance and generalization
 - 🎯 **Unified DPO training** with a single preference dataset across all agents
 - 📏 **Extended context length** from 40k to 64k for more challenging multi-turn tool-use tasks
 

apps/gradio-demo/README.md

@@ -23,7 +23,7 @@ snapshot_download(repo_id="miromind-ai/MiroThinker-v1.5-30B", local_dir="model/M
 
 ### Option 1: SGLang Server (Recommended)
 
-FP8 is a highly efficient 8-bit floating point format that significantly reduces memory usage while maintaining model quality. This approach provides excellent performance for inference workloads on modern GPUs.
+FP8 is a highly efficient 8-bit floating point format that significan'tly reduces memory usage while maintaining model quality. This approach provides excellent performance for inference workloads on modern GPUs.
 
 Please install [SGLang](https://github.com/sgl-project/sglang) first. Then initialize fast inference with FP8 precision:
 

apps/miroflow-agent/benchmarks/evaluators/eval_utils.py

@@ -77,7 +77,7 @@
 
 
 Also note the following things:
-- For grading questions where the gold target is a number, the predicted answer needs to be correct to the last significant figure in the gold answer. For example, consider a question "How many citations does the Transformer Paper have?" with gold target "120k". 
+- For grading questions where the gold target is a number, the predicted answer needs to be correct to the last significan't figure in the gold answer. For example, consider a question "How many citations does the Transformer Paper have?" with gold target "120k". 
     - Predicted answers "120k", "124k", and 115k" are all CORRECT. 
     - Predicted answers "100k" and "113k" are INCORRECT. 
     - Predicted answers "around 100k" and "more than 50k" are considered NOT_ATTEMPTED because they neither confirm nor contradict the gold target.