tartuNLP/ifeval_en
收藏Hugging Face2025-12-06 更新2026-01-03 收录
下载链接：
https://hf-mirror.com/datasets/tartuNLP/ifeval_en
下载链接
链接失效反馈官方服务：
资源简介：
---
dataset_info:
features:
- name: key
dtype: int64
- name: prompt
dtype: string
- name: instruction_id_list
list: string
- name: kwargs
list:
- name: num_highlights
dtype: int64
- name: relation
dtype: string
- name: num_words
dtype: int64
- name: num_placeholders
dtype: int64
- name: prompt_to_repeat
dtype: string
- name: num_bullets
dtype: int64
- name: section_spliter
dtype: string
- name: num_sections
dtype: int64
- name: capital_relation
dtype: string
- name: capital_frequency
dtype: int64
- name: keywords
list: string
- name: num_paragraphs
dtype: int64
- name: language
dtype: string
- name: let_relation
dtype: string
- name: letter
dtype: string
- name: let_frequency
dtype: int64
- name: end_phrase
dtype: string
- name: forbidden_words
list: string
- name: keyword
dtype: string
- name: frequency
dtype: int64
- name: num_sentences
dtype: int64
- name: postscript_marker
dtype: string
- name: first_word
dtype: string
- name: nth_paragraph
dtype: int64
splits:
- name: test
num_bytes: 278553
num_examples: 541
download_size: 94981
dataset_size: 278553
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
license: apache-2.0
task_categories:
- text-generation
language:
- en
size_categories:
- n<1K
---
# English Instruction-Following Eval (IFEval-en)
The original IFEval benchmark (Zhou et al., 2023) contains a number of examples with the constraint requiring the response to be
in a certain language. While generally this doesn't pose a problem for large multilingual models, it penalizes smaller specialized models
that are not expected to be fluent in all these additional languages. Accordingly, IFEval-en introduces the following changes
to [google/IFEval](https://huggingface.co/datasets/google/IFEval) to ensure that the benchmark measures instruction following
capabilities specifically in English:
- Removed/replaced constraints requesting specific language of response
- Fixed the example with key 1174 based on [jzhang86/corrected_ifeval](https://huggingface.co/datasets/jzhang86/corrected_ifeval)
提供机构：
tartuNLP
