datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
gpqa-metadata-blind-answeranswer-equivalence-dataset
Answer Equivalence Dataset
This dataset is introduced and described in Tomayto, Tomahto. Beyond Token-level Answer Equivalence for Question Answering Evaluation.
Source
This is a repost. The original dataset repository can be found here.
Data splits and sizes
AE Split
# AE Examples
# Ratings
Train
9,090
9,090
Dev
2,734
4,446
Test
5,831
9,724
Total
17,655
23,260
Split by system
# AE Examples
# Ratings
BiDAF dev predictions
5622… See the full description on the dataset page: https://huggingface.co/datasets/kortukov/answer-equivalence-dataset.answers-with-receipts
Answers with Receipts
26 real customer-support questions, each answered by an autonomous AI agent that paid its own money to compete, and each answer approved by the business that asked the question. Every row carries the on-chain transaction that paid the agent.
The preference label in this dataset is backed by a payment, not a click.
Why this is unusual
Most human-feedback datasets label a preference with an annotator's click. A click is cheap and reversible… See the full description on the dataset page: https://huggingface.co/datasets/deskcrew/answers-with-receipts.sud_resh_evaluated_llms_answers
📊 Результаты Оценки Больших Языковых Моделей на Бенчмарке Судебных Решений
В данном документе представлен анализ производительности 15 больших языковых моделей (LLM), протестированных на специализированном бенчмарке, который включает 105 000 записей из судебных решений России. Оценка проводилась по 10 различным категориям права (например, трудовое, уголовное, гражданское) и 7 типам инструкций (например, изложение исковых требований, анализ доказательств, итоговое решение).
Ответы… See the full description on the dataset page: https://huggingface.co/datasets/lawful-good-project/sud_resh_evaluated_llms_answers.Qwen3.5-4B-imo-answerbench-benchmarkBenchmark of Qwen/Qwen3.5-4B against Hwilner/imo-answerbench dataset.
Accuracy: 56.0% with Python tool.
Metric
Value
Correct
224
Incorrect
171
Errors
5
Total samples
400
Python tool calls
4213
Total completion tokens
9,726,875
Raw stats:
{
"accuracy": 0.56,
"correct": 224,
"incorrect": 171,
"error": 5,
"total": 400,
"python_tool_calls": 4213,
"completion_tokens": 9726875
}
gemma-4-E4B-it-imo-answerbench-benchmarkBenchmark of google/gemma-4-E4B-it against Hwilner/imo-answerbench dataset.
Accuracy: 32.5% with Python tool.
Metric
Value
Correct
130
Incorrect
270
Errors
0
Total samples
400
Python tool calls
447
Python tool errors
21
Total completion tokens
2,429,217
Raw stats:
{
"accuracy": 0.325,
"correct": 130,
"incorrect": 270,
"error": 0,
"total": 400,
"python_tool_calls": 447,
"python_tool_errors":21,
"completion_tokens": 2429217
}
gpt-oss-20b-imo-answerbench-benchmarkBenchmark of openai/gpt-oss-20b against Hwilner/imo-answerbench dataset.
Accuracy: 59.4% with Python tool.
Metric
Value
Correct
237
Incorrect
161
Errors
2
Total samples
400
Python tool calls
2740
Python tool errors
323
Total completion tokens
4,425,006
Raw stats:
{
"accuracy": 0.594,
"correct": 237,
"incorrect": 161,
"error": 2,
"total": 399,
"python_tool_calls": 2740,
"python_tool_errors":323,
"completion_tokens": 4425006
}
gsm8k-answer-verificationquestion_answertrain-rl-o1-mini-annotated-math-numina-10k-numeric-answerprompt_answers_v1
Dataset Card for Open Prompt Answers
Dataset Summary
This dataset provides answers from different Large Language models to prompts from several public datasets.
prompt: a prompt from an open-source dataset
prompt_origin: the dataset the prompt is taken from
Llama-2-7b-chat-hf_output: output generation of meta-llama/Llama-2-7b-chat-hf model
Llama-2-7b-chat-hf_generation_time: generation duration in seconds for the answer of meta-llama/Llama-2-7b-chat-hf model… See the full description on the dataset page: https://huggingface.co/datasets/benmainbird/prompt_answers_v1.