datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
lmsys-chat-enLMSYS-USP
LMSYS-USP Dataset
Overview
GitHub repository for exploring the source code and additional resources: https://github.com/wangkevin02/USP
The LMSYS-USP dataset contains high-quality dialogues with inferred user profiles(provide natural descriptions encompassing both objective facts and subjective characteristics), generated through a two-stage profiling pipeline (see our paper for details). The dataset includes a training set (87,882 examples), a validation set (4,626)… See the full description on the dataset page: https://huggingface.co/datasets/wangkevin02/LMSYS-USP.lmsys-chat-1m-smortmodelsonlyThis version of the dataset only has responses from GPT-4, Claude-1, Claude-2, Claude-instant-1, and GPT-3.5-turbo
lmsys-arena-human-preference-winner-43k-unfiltered
lmsys-arena-human-preference-winner-43k-unfiltered
This repository contains a dataset derived from the lmsys/lmsys-arena-human-preference-55k dataset, which is licensed under the Apache 2.0 License.
Dataset Description
The lmsys-arena-human-preference-winner-43k-unfiltered dataset is a collection of 43,000 samples, each containing an instruction (prompt) and an output (winning response) from real-world user and LLM conversations. The dataset is derived from the original… See the full description on the dataset page: https://huggingface.co/datasets/lesserfield/lmsys-arena-human-preference-winner-43k-unfiltered.lmsys-chat-1m-filtered
Dataset Card for natong19/lmsys-chat-1m-filtered
Filtered version of lmsys/lmsys-chat-1m, a collection of one million real-world conversations with various LLMs.
Data cleaning process inspired by OpenLeecher/lmsys_chat_1m_clean.
Overview of filtering process:
1. Filtering REDACTED Entries
Entries that were labeled as REDACTED due to containing Personally Identifiable Information (PII) were removed.
1000000 samples -> 733740 samples
2. Format validation… See the full description on the dataset page: https://huggingface.co/datasets/natong19/lmsys-chat-1m-filtered.lmsys-chat-1m-jsonify-v2ytz20-LMSYS-Chat-GPT-5-Chat-Response
ytz20/LMSYS-Chat-GPT-5-Chat-Response Dataset
This is a reformatted, unofficial version of ytz20/LMSYS-Chat-GPT-5-Chat-Response
According to the original authors:
This dataset is an extension of the LMSYS-Chat-1M-Clean corpus, specifically curated by collecting high-quality, non-refusal responses from the GPT-5-Chat API.
Modifications in this version:
The "content" and "teacher_response" columns have been processed into the "input" and "output" columns in this dataset.
An… See the full description on the dataset page: https://huggingface.co/datasets/agentlans/ytz20-LMSYS-Chat-GPT-5-Chat-Response.lmsys-chat-outputslmsys_chat_generated_muselmsys__vicuna-7b-v1.5-details
Dataset Card for Evaluation run of lmsys/vicuna-7b-v1.5
Dataset automatically created during the evaluation run of model lmsys/vicuna-7b-v1.5
The dataset is composed of 81 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 8 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An additional… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/lmsys__vicuna-7b-v1.5-details.lmsys-chat-1m-4keyslmsys__vicuna-7b-v1.3-details
Dataset Card for Evaluation run of lmsys/vicuna-7b-v1.3
Dataset automatically created during the evaluation run of model lmsys/vicuna-7b-v1.3
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An additional… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/lmsys__vicuna-7b-v1.3-details.lmsys_lmsys-chat-1m-gemini-2.0-flash-thinking-exp-1219-CustomShareGPT
lmsys_lmsys-chat-1m-gemini-2.0-flash-thinking-exp-1219-CustomShareGPT
PJMixers-Dev/musab-mk_lmsys-chat-1m_deduped-prompts with responses generated with gemini-2.0-flash-thinking-exp-1219.
Generation Details
If BlockedPromptException, StopCandidateException, or InvalidArgument was returned, the sample was skipped.
If ["candidates"][0]["safety_ratings"] == "SAFETY" the sample was skipped.
If ["candidates"][0]["finish_reason"] != 1 the sample was skipped.
model =… See the full description on the dataset page: https://huggingface.co/datasets/PJMixers-Dev/lmsys_lmsys-chat-1m-gemini-2.0-flash-thinking-exp-1219-CustomShareGPT.lmsys__vicuna-13b-v1.3-details
Dataset Card for Evaluation run of lmsys/vicuna-13b-v1.3
Dataset automatically created during the evaluation run of model lmsys/vicuna-13b-v1.3
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An additional… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/lmsys__vicuna-13b-v1.3-details.Lmsys-1M-TNSA-JSONlmsys_chat_80_EE_multi_turn
Dataset Description
This dataset consists of 80 prompts translated into Estonian from the LMSYS-Chat-1M dataset. The prompts are drawn from four categories: planning and scheduling (planeerimine), specific format writing (kirjutamine), mathematics (matemaatika), and logical reasoning (arutlemine).
Thirty prompts in this dataset are previous translations from smugri-mt-bench and are included either verbatim or with minor modifications.
Prompt Structure
All prompts are… See the full description on the dataset page: https://huggingface.co/datasets/mariiojastu/lmsys_chat_80_EE_multi_turn.gemma-3-12b-it-lmsys-onpolicy-rollouts
On-policy chat rollouts: google/gemma-3-12b-it on LMSYS-Chat-1M prompts
Each row is a first-user-turn prompt sampled from
lmsys/lmsys-chat-1m and a
response generated on-policy by google/gemma-3-12b-it with vLLM (do_sample,
temperature 0.7, top_p 1.0, max_new_tokens 768, seed 42). 24,991 rows. Built to match
GemmaScope 2's instruction-tuned SAE training distribution (real model rollouts) for a
short KL+MSE ("end-to-end") finetune of the released GemmaScope-2 residual SAE.… See the full description on the dataset page: https://huggingface.co/datasets/iarcuschin/gemma-3-12b-it-lmsys-onpolicy-rollouts.lmsys_persian_translationAugmentoolkit-LMsys-800k-Thoughtslmsys-chat-tiny-20kLMSYS-24Kmusab-mk_lmsys-chat-1m_deduped-promptslmsys-arena-human-preference-5k-reformattedlmsys-800ktokens-basicfilteringThis dataset https://huggingface.co/datasets/OpenLeecher/lmsys_chat_1m_clean
But with some basic string matching filtering for refusals applied.
This should improve the generalist capability of your model at the expense of making it a bit censored, since I didn't get it all.
Could use a more sophisticated filtering pipeline for this. I might build a "filter out refusals" pipeline actually. Might take an hour or two.
lmsys_chat_1m_clean_openleacherlmsys-chat-1m-prompts
Prompts from "LMSYS-Chat-1M: A Large-Scale Real-World LLM Conversation Dataset"
This dataset features unique first-turn prompts in English, drawn from the LMSYS-Chat-1M dataset.
It captures how real users begin conversations with large language models (LLMs) in everyday scenarios.
Due to licensing restrictions, this dataset is access-controlled, so this public repository is gated.
[!IMPORTANT]Please do not request access here, as such requests will not be approved.
If you’re… See the full description on the dataset page: https://huggingface.co/datasets/agentlans/lmsys-chat-1m-prompts.lmsys-self-identification-newlmsys_chat_1m_clean_openleacher