Update README.md

README.md

@@ -606,7 +606,7 @@ tags:
 
 ## Dataset Summary
 
-LIBRA (Long Input Benchmark for Russian Analysis) is designed to evaluate the capabilities of large language models (LLMs) in understanding and processing long texts in Russian. This benchmark includes 21 datasets adapted for different tasks and complexities. The tasks are divided into four complexity groups and allow evaluation across various context lengths ranging from 4,000 up to 128,000 tokens.
+LIBRA (Long Input Benchmark for Russian Analysis) is designed to evaluate the capabilities of large language models (LLMs) in understanding and processing long texts in Russian. This benchmark includes 21 datasets adapted for different tasks and complexities. The tasks are divided into four complexity groups and allow evaluation across various context lengths ranging from 4k up to 128k tokens.
 
 ## Dataset Structure