datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
ErrorAnalysis
AnchorSR Error Analysis
500道题，严格沿 failure_cases_500.json 的文件顺序排列，每题对照三个SFT模型。
推荐从第001题开始，点击“下一题”逐题阅读。
也可使用本页上方 Dataset Viewer，一行就是一题：图片、问题、标准答案、三个模型完整输出。
Q-Spatial 150题，SpatialRGPT 175题，VSI 175题（仅尺寸、距离，不含面积）。
全部500题：每个模型均提供原图和标注图，共3000张图；O编号和帧号来自模型声明。
原图与标注图使用相同源帧和拼图顺序。无效框/帧号或无声明会注明，不补造；此时标注页可能没有框。
原图指未添加模型框的网页展示副本，经过等比例缩放与JPEG编码，并非原始文件字节；SpatialRGPT原有区域标记保留。
视频只展示可绘制对象涉及帧，无有效框时展示第1帧，非完整视频。原图/标注图使用相同帧。
原生输出完整保留，包括循环、截断和格式错误；未修改答案或重新评分。
所有模型均为SFT，不是baseline。至少一个模型在该题失败，其他模型可能答对。… See the full description on the dataset page: https://huggingface.co/datasets/AnchorSR/ErrorAnalysis.ag_news_error_analysisThis is a dataset created for testing purposes in the context of this tutorial: https://rubrix.readthedocs.io/en/master/tutorials/08-error_analysis_using_loss.html
You can find more details on section 5. of the tutorial and the corresponding dataset with corrected labels at https://huggingface.co/datasets/Recognai/ag_news_corrected_labels
granite_4.0_h_small_FP8_test_detoxificability_annotation_error_analysisllm-error-analysis
LLM Error Analysis Dataset
This dataset contains evaluation examples for testing mathematical and logical reasoning capabilities of language models.
Dataset Overview
Number of examples: 11
Fields: input, type, expected_output, model_output
Types included: arithmetic, long_addition, date_knowledge, ambiguous_riddle, ordering, directional_reasoning, pattern_recognition, instruction_following, framing_bias, causal_vs_correlation
Model Tested
Model:… See the full description on the dataset page: https://huggingface.co/datasets/ParamTh/llm-error-analysis.LocoOperator-4B-error-analysis
LocoOperator-4B Error Analysis Dataset README
This dataset contains examples of identified errors and unexpected behaviors from the LocoreMind/LocoOperator-4B model, collected during an evaluation phase. It is intended to serve as a basis for fine-tuning the model to improve its performance in specific areas.
Model Tested
Model Name: LocoreMind/LocoOperator-4B
Hugging Face Model Page: https://huggingface.co/LocoreMind/LocoOperator-4B
Model Loading Code
The… See the full description on the dataset page: https://huggingface.co/datasets/EngrIBGIT/LocoOperator-4B-error-analysis.Qwen3-30B-A3B-Instruct-2507-FP8_detoxificability_annotation_error_analysisdetox_eval_aggregated_quality_annotation_error_analysisllm-error-analysis-dataset
Model Evaluation and Fine-Tuning Recommendation
For this experiment, I used the Nanbeige4-3B-Base model from Hugging Face and performed prompt testing across multiple domains.
Model Setup
The following code was used to load the model and generate responses:
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
model_name = "Nanbeige/Nanbeige4-3B-Base"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(… See the full description on the dataset page: https://huggingface.co/datasets/AnindaRoy/llm-error-analysis-dataset.