Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
284
datasets available to search
ShareScore release 0.9.0
Dataset results
284 results for “Reasonableness”
MathChat: Benchmarking Mathematical Reasoning and Instruction Following in Multi-Turn Interactions
<div> <h3>1. follow_up.jsonl</h3> <a href="https://github.com/Zhenwen-NLP/MathChat#1-follow_upjsonl"></a></div> <p>This file contains entries that facilitate follow-up questioning. Each line consists of three keys:</p> <ul> <li><strong>question</strong>: Sourced from the GSM8k testing set.</li> <li><strong>answer</strong>: Corresponding answer from the GSM8k testing set.</li> <li><strong>followup</strong>: Includes two rounds of follow-up questions and reference answers, formatted as a conversation between a user (A:) and an assistant (B:).</li> </ul> <div> <h3>2. error_correction.jsonl</h3> <a href="https://github.com/Zhenwen-NLP/MathChat#2-error_correctionjsonl"></a></div> <p>This file is designed for error correction tasks. Each line consists of three keys:</p> <ul> <li><strong>question</strong>: Sourced from the GSM8k testing set.</li> <li><strong>answer</strong>: Corresponding answer from the GSM8k testing set.</li> <li><strong>error_correction</strong>: Contains a conversation between a user (A:) and an assistant (B:), which includes the original question, an incorrect answer, and the process of correcting the error.</li> </ul> <div> <h3>3. error_analysis.jsonl</h3> <a href="https://github.com/Zhenwen-NLP/MathChat#3-error_analysisjsonl"></a></div> <p>This file also focuses on error correction but employs a different prompt strategy. Each line consists of three keys:</p> <ul> <li><strong>question</strong>: Sourced from the GSM8k testing set.</li> <li><strong>answer</strong>: Corresponding answer from the GSM8k testing set.</li> <li><strong>error_analysis</strong>: Includes a conversation between a user (A:) and an assistant (B:), where the model is prompted to independently determine the correctness of the answer without being explicitly told.</li> </ul> <div> <h3>4. p2p_generation.jsonl</h3> <a href="https://github.com/Zhenwen-NLP/MathChat#4-p2p_generationjsonl"></a></div> <p>This file contains entries for problem generation tasks. Each line consists of three keys:</p> <ul> <li><strong>question</strong>: Sourced from the GSM8k testing set.</li> <li><strong>answer</strong>: Corresponding answer from the GSM8k testing set.</li> <li><strong>new_problem</strong>: A new problem generated by GPT-4 to serve as a reference answer.</li> </ul>
Artifact for "Verifying the Option Type with Rely-Guarantee Reasoning"
<p>This repository contains the dataset and scripts to reproduce Table 1 and Table 2 in the paper "Verifying the Option Type with Rely-Guarantee Reasoning".</p> <h2>Requirements</h2> <p>Please see <code>REQUIREMENTS.md</code> for the same information, but listed separately from this README file.</p> <p>Our scripts are designed for a Unix environment (Linux, macOS, WSL, etc.). They will not work under a Windows command shell.</p> <p>Our scripts require the following tools:</p> <ul> <li> <p><a href="https://www.python.org/downloads/">Python 3</a>, version 3.9.6 or later.</p> </li> <li> <p>GNU <code>grep</code>. For macOS, install it as <code>ggrep</code> via the command <code>brew install grep</code>.</p> <ul> <li>If not already installed, <code>brew</code> can be installed via the instructions<a href="https://docs.brew.sh/Installation">here</a></li> </ul> </li> </ul> <h2>Subject programs</h2> <p>File <code>dataset.txt</code> lists the subject programs.</p> <p>The subject programs are stored under the following directories:</p> <ul> <li><code>optional-paper-base-repos</code>: all the subject programs with their build files modified to run the Optional Checker.</li> <li><code>optional-paper-annotated-repos</code>: all the subject programs, with<code>@SuppressWarnings</code> for each true positive and false positive issued by the Optional Checker (scripts count these). It also contains the Optional Checker type qualifiers that we wrote.</li> <li><code>optional-paper-intellij-repos</code>: all the subject programs with <code>@SuppressWarnings</code> for each IntelliJ warning (so they can be programmatically counted).</li> <li><code>optional-paper-errorprone-repos</code>: all the subject programs with <code>@SuppressWarnings</code> for each Error Prone warning.</li> <li><code>optional-paper-spotbugs-repos</code>: all the subject programs with <code>@SuppressWarnings</code> for each SpotBugs warning.</li> </ul> <h2>Scripts</h2> <p>Looking to reproduce <em>all</em> the data in our tables? See <a href="reproduce-data"><code>reproduce-data</code></a>.</p> <p>This repository contains the following scripts:</p> <ul> <li><a href="reproduce-data"><code>reproduce-data</code></a>: this script reproduces all the data for our tables.</li> <li><a href="compute-precision-recall-annotations"><code>compute-precision-recall-annotations</code></a>: this script generates Table 1: it computes the values for precision, recall, and the number of human-written, machine-checked annotations used by each tool.</li> <li><a href="count-style-violations"><code>count-style-violations</code></a>: this script generates Table 2: it reports the number of style violations detected by each tool.</li> <li><a href="mygrep.py"><code>mygrep.py</code></a>: this is a Python utility used by other scripts as a thin wrapper around <code>grep</code>. You need not use it directly.</li> </ul> <p>All these scripts may be executed from the root of this directory. For example, to generate Table 1:</p> <pre><code>% ./compute-precision-recall-annotations </code></pre> <h2>Resulting Data</h2> <p>Each script produces the <em>rows</em> of the dataset, excluding the column titles. Below is a brief description of each output <code>.tex</code> file:</p> <ul> <li><code>eval-statistics.tex</code>: generated by the <code>compute-precision-recall-annotations</code> script, maps to Table 1 in our paper. The columns of the data are, in order from left-to-right: <ul> <li>Tool name</li> <li>Number of true positives detected by the tool</li> <li>Number of false positives detected by the tool</li> <li>Precision</li> <li>Recall</li> <li>Total number of machine-verified annotations written as a specification across all subject programs.</li> </ul> </li> <li><code>style-violations.tex</code>: generated by the <code>count-style-violations</code>script, maps to Table 2 in our paper. The columns are the data are, in order from left-to-right: <ul> <li>Tool name</li> <li>Number of violations of style rule 1</li> <li>Number of violations of style rule 3</li> <li>Number of violations of style rule 4</li> <li>Number of violations of style rule 5</li> <li>Number of violations of style rule 6.a</li> <li>Number of violations of style rule 6.b</li> <li>Number of violations of style rule 6.c</li> <li>Number of violations of style rule 7</li> </ul> </li> </ul> <h2>Implementation</h2> <p>The implementation of the Optional Checker appears in the <code>implementation</code> directory. This folder contains the following subdirectories:</p> <ul> <li><code>optional</code>: the source code for the Optional Checker's verification logic for the Optional type system. <ul> <li><code>optional/qualifiers</code>: the definitions for each type qualifier from the Optional type system that may be used with the Optional Checker.</li> </ul> </li> <li><code>nonempty</code>: the source code for the Optional Checker's verification logic for the Non-Empty type system. <ul> <li><code>nonempty/qualifiers</code>: the definitions for each type qualifier from the Non-Empty type system that may be used with the Optional Checker.</li> </ul> </li> </ul>
REsolved ALMA and SMA Observations of Nearby Stars (REASONS): Observation Logs
<p>Observation Details for millimetre-wavelength data gathered and modelled as part of the REsolved ALMA and SMA Observations of Nearby Stars (REASONS) paper (same DOI as this table). For the processed data and modelling results, see https://doi.org/10.5281/zenodo.12583396.</p> <ul> <li>Table A.1 refers to targets observed as part of the REASONS <em>Observational</em> ALMA program</li> <li>Table A.2 refers to targets observed as part of the REASONS <em>Observational</em> SMA program</li> <li>Table A.3 refers to targets previously observed whose raw data was retrieved from the ALMA archive and reanalysed as part of REASONS.</li> </ul>
Supplemental data to paper "An Extensible Logic Embedding Tool for Lightweight Non-Classical Reasoning"
<p>These files are supplementary material to the paper<br> An Extensible Logic Embedding Tool for Lightweight Non-Classical Reasoning<br> Alexander Steen<br> accepted at the Eighth Workshop on Practical Aspects of Automated Reasoning (PAAR 2022), 2022.</p> <p>Files ex1.p, ex2.p and ex3.p correspond to Examples 1, 2, and 3 from the paper in section 5.<br> The *.embed.p files are their embedded variants in classical HOL (THF).</p> <p>In a later update, according to the paper [1], the syntax of NXF and NHF was slightly updated to include<br> an "@"-sign between the non-classical operator {...} and its arguments (...), even in NXF. As a consequence,<br> the examples in this archive have been updated accordingly.</p> <p>Alexander Steen, <alexander.steen@uni-greifswald.de></p> <p> </p> <p>[1] Alexander Steen, David Fuenmayor, Tobias Gleißner, Geoff Sutcliffe, Christoph Benzmüller: Automated Reasoning in Non-classical Logics in the TPTP World. PAAR@IJCAR 2022. Available at https://ceur-ws.org/Vol-3201/paper11.pdf</p>
Simple Multimodal Algorithmic Reasoning Task Dataset (SMART-101)
<p><strong>Introduction</strong></p> <p>Recent times have witnessed an increasing number of applications of deep neural networks towards solving tasks that require superior cognitive abilities, e.g., playing Go, generating art, ChatGPT, etc. Such a dramatic progress raises the question: how generalizable are neural networks in solving problems that demand broad skills? To answer this question, we propose SMART: a Simple Multimodal Algorithmic Reasoning Task (and the associated SMART-101 dataset) for evaluating the abstraction, deduction, and generalization abilities of neural networks in solving visuo-linguistic puzzles designed specifically for children of younger age (6--8). Our dataset consists of 101 unique puzzles; each puzzle comprises a picture and a question, and their solution needs a mix of several elementary skills, including pattern recognition, algebra, and spatial reasoning, among others. To train deep neural networks, we programmatically augment each puzzle to 2,000 new instances; each instance varied in appearance, associated natural language question, and its solution. To foster research and make progress in the quest for artificial general intelligence, we are publicly releasing our SMART-101 dataset, consisting of the full set of programmatically-generated instances of 101 puzzles and their solutions.</p> <p>The dataset was introduced in our paper <a href="https://arxiv.org/pdf/2212.09993.pdf">Are Deep Neural Networks SMARTer than Second Graders?</a> by Anoop Cherian, Kuan-Chuan Peng, Suhas Lohit, Kevin A. Smith, and Joshua B. Tenenbaum, CVPR 2023</p> <p>Files in the unzipped folder:</p> <ol> <li>./README.md: This Markdown file</li> <li>./SMART101-Data: Folder containing all the puzzle data. See below for details.</li> <li>./puzzle_type_info.csv: Puzzle categorization (into 8 skill classes).</li> </ol> <p><strong>Dataset Organization</strong></p> <p>The dataset consists of `101` folders (numbered from 1-101); each folder corresponds to one distinct puzzle (root puzzle). There are 2000 puzzle instances programmatically created for each root puzzle, numbered from 1-2000. Every root puzzle index (in [1,101]) folder contains: (i) `img/` and (ii) `puzzle_<index>.csv`. The folder `img/` is the location where the puzzle instance images are stored, and `puzzle_<index>.csv` the non-image part of a puzzle. Specifically, a row of `puzzle_<index>.csv` is the following tuple: `<id, Question, image, A, B, C, D, E, Answer>`, where `id` is the puzzle instance id (in [1,2000]), `Question` is the puzzle question associated with the instance, `image` is the name of the image (in `img/` folder) corresponding to this instance `id`, `A, B, C, D, E` are the five answer candidates, and `Answer` is the answer to the question. </p> <p><strong>At a Glance</strong></p> <ul> <li>The size of the unzipped dataset is ~12GB. </li> <li>The dataset consists of `101` folders (numbered from 1-101); each folder corresponds to one distinct puzzle (root puzzle). </li> <li>There are 2000 puzzle instances programmatically created for each root puzzle, numbered from 1-2000. </li> <li>Every root puzzle index (in [1,101]) folder contains: (i) `img/` and (ii) `puzzle_<index>.csv`. </li> <li>The folder `img/` is the location where the puzzle instance images are stored, and `puzzle_<index>.csv` contains the non-image part of a puzzle. Specifically, a row of `puzzle_<index>.csv` is the following tuple: `<id, Question, image, A, B, C, D, E, Answer>`, where `id` is the puzzle instance id (in [1,2000]), `Question` is the puzzle question associated with the instance, `image` is the name of the image (in `img/` folder) corresponding to this instance `id`, `A, B, C, D, E` are the five answer candidates, and `Answer` is the correct answer to the question. </li> </ul> <p><strong>Other Details</strong><br> In our paper <a href="https://arxiv.org/pdf/2212.09993.pdf">Are Deep Neural Networks SMARTer than Second Graders?</a>, we provide four different dataset splits for evaluation: (i) Instance Split (IS), (ii) Answer Split (AS), (iii) Puzzle Split (PS), and (iv) Few-shot Split (FS). Below, we provide the details of each split to make fair comparisons to the results reported in our paper. </p> <p><em>Puzzle Split (PS)</em><br> We use the following root puzzle ids as the `Train` and `Test` sets. </p> <table> <thead> <tr> <th scope="col">Split</th> <th scope="col">Root Puzzle Id Sets</th> </tr> </thead> <tbody> <tr> <td>`Test`</td> <td>{ 94,95, 96, 97, 98, 99, 101, 61,62, 65, 66,67, 69, 70, 71,72,73,74,75,76,77}</td> </tr> <tr> <td>`Train`</td> <td>{1,2,...,101} \ Test</td> </tr> </tbody> </table> <p>Evaluation is done on all the `Test` puzzles and their accuracies averaged. For the 'Test' puzzles, we use the instance indices 1701-2000 in the evaluation.</p> <p><em>Few-shot Split (FS)</em></p> <p>We randomly select `k` number of instances from the `Test` sets (that are used in the PS split above) for training in FS split (e.g., `k=100`). These `k` few-shot samples are taken from instance indices 1-1600 of the respective puzzles and evaluation is conducted on all instance ids from 1701-2000.</p> <p><em>Instance Split (IS)</em></p> <p>We split the instances under every root puzzle as: Train = 1-1600, Val = 1601-1700, Test = 1701-2000. We train the neural network models using the `Train` split puzzle instances from all the root puzzles together and evaluate on the `Test` split of all puzzles.</p> <p><em>Answer Split (AS)</em></p> <p>We find the median answer value among all the 2000 instances for every root puzzle and only use this set of the respective instances (with the median answer value) as the `Test` set for evaluation (this set is excluded from the training of the neural networks).</p> <p><em>Puzzle Categorization</em></p> <p>Please see puzzle_type_info.csv for details on the categorization of the puzzles into eight classes, namely (i) counting, (ii) logic, (iii) measure, (iv) spatial, (v) arithmetic, (vi) algebra, (vii) pattern finding, and (viii) path tracing.</p> <p><strong>Other Resources</strong></p> <p>PyTorch code for using the dataset to train deep neural networks is available <a href="https://www.merl.com/publications/TR2023-014">here</a>.</p> <p><strong>Contact</strong><br> Anoop Cherian (cherian@merl.com), Kuan-Chuan Peng (kpeng@merl.com), or Suhas Lohit (slohit@merl.com)</p> <p><br> <strong>Citation</strong><br> If you use the SMART-101 dataset in your research, please cite our paper:</p> <pre><code>@article{cherian2022deep, title={Are Deep Neural Networks SMARTer than Second Graders?}, author={Cherian, Anoop and Peng, Kuan-Chuan and Lohit, Suhas and Smith, Kevin and Tenenbaum, Joshua B}, journal={arXiv preprint arXiv:2212.09993}, year={2022} }</code></pre> <p><strong>Copyright and Licenses</strong></p> <p>The SMART-101 dataset is released under `CC-BY-SA-4.0`.</p> <pre><code>Created by Mitsubishi Electric Research Laboratories (MERL), 2022-2023 SPDX-License-Identifier: CC-BY-SA-4.0 </code></pre> <p> </p>
CARP: Evaluating and Improving Tool-Augmented Computation-Intensive Math Reasoning
<p>The CARP dataset consists of 4,886 middle school computation-intensive algebra problems, and each problem is associated with a natural language solution and an annotated EFG.<br> Our annotated EFG explicitly depicts the step-by-step reasoning process of a math problem in a readable and concise format.</p>
Data from: Two complementary AI approaches for predicting UMLS semantic group assignment: heuristic reasoning and deep learning
<p><strong>Objective</strong>: Use heuristic, deep learning (DL), and hybrid AI methods to predict semantic group (SG) assignments for new UMLS Metathesaurus atoms, with target accuracy ≥ 95%.</p> <p><strong>Materials and Methods</strong>: We used train-test datasets from successive 2020AA-2022AB UMLS Metathesaurus releases. Our heuristic "waterfall" approach employed a sequence of seven different SG prediction methods. Atoms not qualifying for a method were passed on to the next method. The DL approach generated BioWordVec and SapBERT embeddings for atom names, BioWordVec embeddings for source vocabulary names, and BioWordVec embeddings for atom names of the second-to-top nodes of an atom's source hierarchy. We fed a concatenation of the four embeddings into a fully connected multi-layer neural network with an output layer of 15 nodes (one for each SG). Both methods were capable of estimating the probability that their predicted SG for an atom would be correct. We developed two hybrid SG prediction methods combining the strengths of heuristic and DL methods.</p> <p><strong>Results</strong>: The heuristic waterfall approach accurately predicted 94.3% of SGs for 1,563,692 new unseen atoms. The DL accuracy on the same dataset was also 94.3%. The hybrid approaches achieved an average accuracy of 96.5%.</p> <p><strong>Conclusion</strong>: Our study demonstrated that AI methods can predict SG assignments for new UMLS atoms with sufficient accuracy to be potentially useful as an intermediate step in the time-consuming task of assigning new atoms to UMLS concepts (CUIs). We showed that for SG prediction, combining heuristic methods and DL methods can produce better results than either alone.</p>
ThoughtSource: A central hub for large language model reasoning data (dataset snapshot)
<p><strong>ThoughtSource is a meta-dataset and software library for chain-of-thought reasoning in large language models (LLMs). </strong></p> <p><strong>This repository contains a snapshot of the openly available ThoughtSource datasets.</strong></p>
The Metadata of "Are the Lists of Questionable Journals Reasonable: A Case Study of Early Warning Journal List"
<p>The metadata of the article (<em>Are the Lists of Questionable Journals Reasonable: A Case Study of Early Warning Journal List</em>).</p>
TG-CSR: Theoretically-Grounded Commonsense Reasoning Benchmark
<p>Achieving machine common sense has been a longstanding problem within Artificial Intelligence. Thus far, benchmark data sets that are grounded in a theory of common sense and can be used to conduct rigorous, semantic evaluations of common sense reasoning (CSR) systems have been lacking. One expectation of the AI community is that neuro-symbolic reasoners can help bridge this gap towards more dependable systems with common sense. We propose a novel benchmark, called <em>Theoretically Grounded common sense Reasoning (TG-CSR)</em>, modeled as a set of question-answering instances, with each instance grounded in a semantic category of common sense, such as space, time, and emotions. The benchmark is few-shot i.e., only a few training and validation examples are provided in the public release to avoid the possibility of overfitting. Results from recent evaluations suggest that TG-CSR is challenging even for state-of-the-art statistical models. Due to its semantic rigor, this benchmark can be used to evaluate the common sense reasoning capabilities of neuro-symbolic systems.</p>
Data for Reserach titled "Exploring Scientific Reasoning and Genetics Self-efficacy as Predictors of Secondary School Students' Achievement in Genetics Concepts"
<p>This dataset is for the research article titled Exploring Scientific Reasoning and Genetics Self-efficacy as Predictors of Secondary School Students’ Achievement in Genetics Concepts. The data were collected using three instruments (Lawson’s Classroom Test of Scientific Reasoning (r = 0.76), adopted from the Lawson's Classroom Test of Scientific Reasoning designed by Lawson in 2000; Self-efficacy in Genetics Concepts Scale (@ = 0.88),adapted from Schwarzer and Jerusalem (1995) General Self Efficacy Scale and Biology Self-Efficacy developed by Baldwin, Ebert-May and Burns (1999); and Students’ Achievement Test in Genetics Concepts (r = 0.78), self-designed by the athors) from 336 Secondary School Two (SS2) biology students selected from nine secondary schools randomly selected from three local government areas in Oyo State. SPSS version 26 was employed as the statitsical software, while data collected were subjected to mean, standard deviation, percentages, ANOVA and Multiple regression analysis at 0.05 level of significance.</p>
Microservice ChatGPT Reasoning
<p>This dataset contains:</p> <p>- The source code data extraction from TrainTicket testbench V0.1.0.</p> <p>- The PO-CCG (Persistence Operation aware Component Call Graph) data extraction of TrainTicket V0.1.0.</p> <p>- The questions/ answers and the evaluation of ChatGPT answers.</p>
Data for paper on inductive reasoning in large language models
<p>This dataset contains all data that was used for:</p> <p>Han, S. J., Ransom, K. J., Perfors, A. & Kemp, C. (2023). Inductive reasoning in humans and large language models. <em>Cognitive Systems Research</em>.</p> <p>The code for this project can be found <a href="https://github.com/S-J-HAN/InductiveReasoningInLargeLanguageModels">here</a>.</p> <p> </p>
Code and data for the paper "Automated reasoning for proving non-orderability of groups"
<p>Code and data for the paper titled "Automated reasoning for proving non-orderability of groups". Containing the input files, output files and proofs obtained by Prover9; the input files, output files and models obtained by Mace4; and a Python 3 script to generate this data set.</p>
Supplementary material for: "Missing nurses cause missed care: is that it? Non-Trivial Configurations of Reasons Associated with Missed Care in Austrian hospitals – A qualitative comparative analysis"
<p>Dataset with calibrated data to reproduce the Qualitative Comparative Analysis in our article. </p><p>This dataset was generated by Ana Cartaxo. It contains 81 variables with 401 observations (complete data), which were included in the MISSCARE-Austria Study and were generated using the revised MISSCARE Austria instrument (Cartaxo et al., 2022). The original data were calibrated as described in the article in preparation of performing Qualitative Comparative Analysis – variables regarding contextual factors (Nurse Characteristics, Unit Characteristics, Hospital Characteristics), reasons for missed nursing care (Demand for patient care, Relationship and communication factors, Labor resources allocation, Material resources allocation) and outcomes of missed nursing care are included in the csv. file. </p><p>The R Script (R-Code.R) was created by Ana Cartaxo, João Cartaxo und Johannes Bergmann. It contains the different functions and analysis steps to reproduce the Qualitative Comparative Analysis reported on the article above, using the dataset made available ("Calibrated_Data_Set.csv").</p>
Natural Language Annotations for Reasoning about Program Semantics
<p>Natural language annotations about Python statements</p> <p>The dataset is made of pairs of files sharing the prefix of the filename</p> <p>* Annotations are in JSONL format (filenames ending with '_annot'), i.e. JSON objects separated by newline ('\n') characters</p> <p>* Reference source code files are in JSON format. (filenames ending with '_code')</p> <p> </p> <p>Source dataset : Programming Puzzles (Schuster et al. 2021, NeurIPS Dataset and benchmarks track) - MIT License - https://github.com/microsoft/PythonProgrammingPuzzles</p>
Brain and Cognitive Changes After Reasoning Training in Individuals With Mild Cognitive Impairment
ClinicalTrials.gov study NCT02596906. IPD Sharing: NO. Countries: 1. Publications: 1.
Effect of AI Chatbot-Assisted Versus Traditional Case-Based Learning on Clinical Reasoning in Occupational Therapy Students: A Study on Parkinson's Disease
ClinicalTrials.gov study NCT07045077. IPD Sharing: NO. Countries: 1. Publications: 1.
GI-Reasons- A Trial Of GI Safety Of Celecoxib Compared With Non-Selective Nonsteroidal Antiinflammatory Drugs (NSAIDS)
ClinicalTrials.gov study NCT00373685. IPD Sharing: Not stated. Countries: 2. Publications: 1.
A Study of Patients With Primary Open Angle Glaucoma or Ocular Hypertension Switched to Lumigan® UD Monotherapy for Medical Reasons
ClinicalTrials.gov study NCT01853085. IPD Sharing: Not stated. Countries: 1. Publications: 1.
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.