Update README.md
Browse files
README.md
CHANGED
|
@@ -7,6 +7,20 @@ tags:
|
|
| 7 |
- BenchLocalPack
|
| 8 |
- IFEval
|
| 9 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
# Описание
|
| 11 |
Это просто тестовые наборы, для BenchLocal, первый, тестовый набор, это IFEval-mini, набор всего на 90 заданий, на русском, английском, и на смеси двух языков.
|
| 12 |
|
|
|
|
| 7 |
- BenchLocalPack
|
| 8 |
- IFEval
|
| 9 |
---
|
| 10 |
+
|
| 11 |
+
# EN
|
| 12 |
+
|
| 13 |
+
# Description
|
| 14 |
+
These are just test sets for BenchLocal, the first test set is IFEval-mini, a set of only 90 tasks, in Russian, English, and a mixture of the two languages.
|
| 15 |
+
|
| 16 |
+
# IFEval-mini
|
| 17 |
+
The set was compiled by Gemini 3.8 Flash. I conducted a test on this set, reviewed the results, and read the tasks, indicating what was written incorrectly in the tasks and how some of them could be corrected. The main goal in compiling IFEval tasks was to write tasks that are read strictly unambiguously and composed in such a way as to comply with the validator.
|
| 18 |
+
|
| 19 |
+
___
|
| 20 |
+
___
|
| 21 |
+
|
| 22 |
+
# RU
|
| 23 |
+
|
| 24 |
# Описание
|
| 25 |
Это просто тестовые наборы, для BenchLocal, первый, тестовый набор, это IFEval-mini, набор всего на 90 заданий, на русском, английском, и на смеси двух языков.
|
| 26 |
|