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.
6
datasets available to search
ShareScore release 0.9.0
Dataset results
6 results for “Artificial Language”
Spanish Language Benchmark for Artificial Intelligence Models (TELEIA)
<h1>TELEIA Datasets</h1> <p>The dataset contains test questions to evaluate LLMs in Spanish</p> <ul> <li>TELEIA_Cervantes_AVE.csv: test questions on vocabulary and grammatical structures, following the format of the Cervantes AVE exam. The column "question" contains a sentence with a gap that must be filled in. The columns "option_a", "option_b", "option_c", "option_d" and "correct_answer" (A,B,C or D) represent the different possibilities for the gap and the correct one.</li> <li>TELEIA_PCE.csv: test on morphology and semantics resembling the style of the PCE exam, consisting of short questions or sentences to be completed. It is composed of the following columns: "question", "option_a", "option_b", "option_c" and "correct_answer" (A, B or C).</li> <li>TELEIA_SIELE.csv: different texts with questions related to them, based on the reading comprehension task of the SIELE exam. It is composed of the following columns: "question", "option_a", "option_b", "option_c" and "correct_answer" (A, B or C). </li> </ul>
Leaderboard Spanish Language Benchmark for Artificial Intelligence Models (TELEIA)
<h1>TELEIA Datasets Leaderboard</h1> <p>These dataset contains the answers of different LLMs to the <a href="../doi/10.5281/zenodo.12571762">TELEIA (Spanish Language Benchmark for Artificial Intelligence Models)</a> dataset.<br><br><span>LLMs evaluated:</span></p> <ul> <li>Yi-6B-Chat</li> <li>Meta-Llama-3-8B-Instruct</li> <li>Llama-2-7b-chat-hf</li> <li>gemma-7b-it</li> <li>Mistral-7B-Instruct-v0.1</li> <li>occiglot-7b-es-en-instruct</li> <li>GPT3.5</li> <li>GPT4</li> </ul> <p><span>Files:</span></p> <ul> <li><em>TELEIA_Cervantes_AVE_results.xlsx: </em>vocabulary and grammatical structures, following the format of the Cervantes AVE exam</li> <li><em>TELEIA_PCE_results.xlsx: </em>test on morphology and semantics resembling the style of the PCE exam, consisting of short questions or sentences to be completed</li> <li><em>TELEIA_SIELE_results.xlsx: </em>different texts with questions related to them, based on the reading comprehension task of the SIELE exam</li> </ul> <p>Each .xlsx contains a sheet with the results of each model and the following columns:</p> <ul> <li><em>question: </em>question from TELEIA</li> <li><em>option_a:</em> possible answer from TELEIA </li> <li><em>option_b:</em> possible answer from TELEIA<em> </em></li> <li><em>option_c: </em>possible answer from TELEIA <em> </em></li> <li><em>option_d: </em>possible answer from TELEIA<em> </em></li> <li><em>correct_answer:</em> correct answer form TELEIA </li> <li><em>llm_question:</em> complete question made to the LLM </li> <li><em>tokens_in:</em> list of tokens that compound the question </li> <li><em>tokens_in_count:</em> number of tokens that compound the question </li> <li><em>llm_answer:</em> raw answer from the LLM </li> <li><em>llm_answer_filtered:</em> answer in format {A,B,C,D} from the LLM </li> <li><em>tokens_out :</em> list of tokens that compound the raw answer </li> <li><em>tokens_out_count:</em> number of tokens that compound the raw answer </li> <li><em>word_count :</em> number of words that compound the raw answer </li> </ul> <p> </p>
Clinic Waiting Room-based Study of Swahili Language Artificial Intelligence-driven Symptom Assessments in Tanzanian Primary Health Care Facilities
ClinicalTrials.gov study NCT04958577. IPD Sharing: NO. Countries: 1. Publications: 2.
THE IMPORTANCE OF PROGRAMMING LANGUAGES IN THE FIELD OF ARTIFICIAL INTELLIGENCE
Open the record for dataset details and reuse information.
A Free Verbalization Method of Evaluating Sound Design: The Effectiveness of Artificially Intelligent Natural Language Processing Methods and Tools
<p>"Robot" voice sound files. Seventeen sound files were recorded in four formats; raw human voiceover (VO), and three types of robot voice: vocoded voice 1 (“robo”), vocoded voice 2 with music (“kbd”), and a “beep” voice. Each was recorded as 44.1kHz, 24-bit wav files in a professional recording studio. VO was recorded by professional voice actor DB Cooper, who has been the robot voice for several video games, as well as the voice of the DEE BMW internal car AI voice system. Cooper recorded three versions of the emotes on a Sennheiser MKH-416. Professional sound designer pdx Drescher, an expert in robot<br> and interface sound design, created three sets of robot voices from<br> the original voice files. With guidance from one of the authors,<br> pdx was tasked with trying different approaches to turning the VO<br> samples into three different types of robot voice while attempting<br> to maintain the meaning of the original sounds as described in the<br> list above through preserving the prosody/melodic contour of the<br> original. The first set, robo, used some clips from one of pdx’s prior<br> robot voice projects and integrated them to approximate the emo-<br> tional intention of the VO. Clips were re-pitched, manipulated, and<br> modulated using ProTools plugins. For the kbd takes, VO sounds<br> were played into a Shure SM58 microphone. Vocoder patches mod-<br> ified the signal by voice formants, and the pitch was determined<br> by MIDI notes and pitch-bend controllers. Output of the synthe-<br> sizer was then edited with additional synth patches and effects (EQ,<br> modulation, etc.). We made particular use of a plugin called Envy<br> by Cargo Cult, which takes the volume, pitch, and EQ envelopes of<br> one sound (the original VO) and apply them to another sound. This<br> helped make the synth resemble the prosody of the original sound<br> to some degree. The beep sounds underwent a similar development<br> process as the robo takes, but with interface “bleeps and bloops”<br> derived from various sound effects libraries, including the Star Trek<br> LCARS soundset.</p> <p>The following sounds were recorded: 1. Warning calm (“Uh-oh”) 2. Warning alarm (“ah!”) 3. Wrong/<br> error (“rrrrr”) 4. Correct/good (“yay”) 5. Surprise (neutral) (“Oh!”) 6.<br> Surprise (good) (“Oh!”) 7. Surprise (bad) “(ohhh”) 8. Love/adoration<br> (“awww”) 9. Disgust (“ew”) 10. Contempt (“ech”) 11. Guilt (“hmmm”)<br> 12. Confused (“huh?”) 13. Laugh (“ha ha”) 14. Calculating (“hmmm”)<br> 15. Sigh 16. Giggle 17. Pain (“ow ")</p>
Language Acquisition in the Brain and Algorithms: Towards Systematic Monitoring of the Evolution of Semantic Representations in Biological and Artificial Neural Networks
ClinicalTrials.gov study NCT05217043. IPD Sharing: NO. Countries: 1. Publications: 0.
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.