datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
kronos-scripts
Kronos — Open-source coding-model training pipeline
Live status: https://kronos.menustudioai.com · R0 adapter: jaivial/kronos-round0-qwen15coder-lora · Logs: this dataset
Kronos is an experiment in autonomously training an Opus-4.6-comparable coding LLM on €100 of seed capital, driven 24/7 by a self-evolving agent ("brain") running on systemd. Every cycle the brain orients on its own SQLite memory, picks ONE action (research / train / benchmark / report / improve / spend / idle)… See the full description on the dataset page: https://huggingface.co/datasets/jaivial/kronos-scripts.Kronos-Dataset
Kronos-Dataset
It is a collection of various datasets to expand the capabilities of reasoning models in agent tasks, medical reasoning, multilingual thinking, and writing. All of these are unified in a single format:
[
{
"from": "system",
"value": "You are a medical AI assistant with advanced reasoning capabilities. Provide detailed, step-by-step analysis for medical questions."
},
{
"from": "human",
"value": "Given the symptoms of sudden weakness in the left arm and leg, recent… See the full description on the dataset page: https://huggingface.co/datasets/FredyRivera-dev/Kronos-Dataset.kronos-2026-booksmath-sft-mixedkronos-2026-publications
Kronos Fusion Energy — 2026 Publication Series
The complete 2026 design-and-de-risking series for a compact spherical-tokamak breeder (Hyperion)
and its D–³He tandem-mirror burners (Aegis / MetroVolt) — 115 papers, each with its editorial
PDF, a data-&-methods companion, machine-readable frozen anchors, and figures.
Every numerical value is a frozen design-point anchor, independently re-run and stamped in the
Physics De-Risking Register (paper 1.0, DOI 10.5281/zenodo.22645689).… See the full description on the dataset page: https://huggingface.co/datasets/KrononosFE/kronos-2026-publications.T145__KRONOS-8B-V6-details
Dataset Card for Evaluation run of T145/KRONOS-8B-V6
Dataset automatically created during the evaluation run of model T145/KRONOS-8B-V6
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/T145__KRONOS-8B-V6-details.kronos-2026-ai-quantum
Physics-Informed AI and Quantum Technologies for Certifiable Real-Time Control of Compact Fusion Generators
Authors: Ford, P. I.
Summary
A layered digital-twin control architecture holding three technology-readiness tiers apart. A control-barrier safety clamp gives a provable zero-escape guarantee on the reduced-order operating envelope; quantum sensing is near-term diagnostics; quantum computation is an offline calibration tier with no resource crossover this… See the full description on the dataset page: https://huggingface.co/datasets/KrononosFE/kronos-2026-ai-quantum.kronos-2026-2-8-quantum-computing
Quantum Simulation of Fusion Kinetics and Materials, Honestly: 5-Qubit Landau Damping, ADAPT-VQE at Chemical Accuracy, and the Resource Frontier
Authors: Ford, P I
Summary
Quantum computing for fusion, with honest resource estimates. Landau damping reproduced on a five-qubit Hamiltonian-simulation circuit; ADAPT-VQE at chemical accuracy on the candidate fusion-alloy fragments; and a resource frontier separating the tractable linear kinetic operator (early… See the full description on the dataset page: https://huggingface.co/datasets/KrononosFE/kronos-2026-2-8-quantum-computing.kronos-2026-2-28-tensor-network-solver
The Tensor-Network Kinetic Solver - Classical, Deployable Today
Authors: Ford, P. I.
Summary
The kinetic distribution function is highly compressible in a low-rank tensor-network representation, and that yields a practical classical solver, not a storage trick. On a 1D1V BGK test, a matrix-product-state truncation reaches relative-L² error 2×10−⁴ at rank 8 using ~0.19× the dense storage, with the error falling exponentially in rank. Unlike the… See the full description on the dataset page: https://huggingface.co/datasets/KrononosFE/kronos-2026-2-28-tensor-network-solver.kronos-2026-2-37-maturity-gated-actuation
Maturity-Gated Actuation Authority - Trust Earned by Evidence
Authors: Ford, P. I.
Summary
A control-governance principle that makes trust quantitative: the actuation authority a controller or AI is permitted at runtime is a monotone function of the verification-and-validation maturity (PCMM level) of the underlying models. Machine-validated subsystems admit wide, fast authority; reduced-order ones are clamped to reversible actions — an auditable rule where… See the full description on the dataset page: https://huggingface.co/datasets/KrononosFE/kronos-2026-2-37-maturity-gated-actuation.kronos-2026-2-38-mlip-discovery-pipeline
The Materials-Discovery Pipeline - Activation as a First-Class Filter
Authors: Ford, P. I.
Summary
The method behind the low-activation alloy family: a three-stage GPU pipeline — a universal ML interatomic potential (CHGNet) screens thousands of compositions by formation energy, density-functional theory confirms the survivors, and an activation-transport down-select (OpenMC→FISPACT) rejects any whose waste class or decay heat is unacceptable. The novelty is… See the full description on the dataset page: https://huggingface.co/datasets/KrononosFE/kronos-2026-2-38-mlip-discovery-pipeline.adaption-football-tactics-qa
This dataset is a remastered version prepared using Adaption's Adaptive Data platform.
adaption-football_tactics_qa
This instruction-tuning dataset comprises prompt and completion pairs focused on association football and sports analytics. Content covers match analysis, tactical formations, player performance metrics, official rules, and strategic gameplay scenarios. The dataset supports conversational AI training by providing clear, well-reasoned responses ranging from general… See the full description on the dataset page: https://huggingface.co/datasets/kronos17/adaption-football-tactics-qa.kronos-2026-2-7-cbf-safety-clamp
A Control-Barrier-Certified Zero-Escape Safety Clamp for a Compact Spherical-Tokamak Breeder
Authors: Ford, P I
Summary
A deterministic, control-barrier-certified safety clamp providing a provable zero-escape guarantee for the compact spherical-tokamak breeder operating envelope. A robust, disturbance-tightened control-barrier-function filter with a closed-form separable solution wraps an arbitrary controller: zero of 5000 Monte-Carlo trajectories escape the safe… See the full description on the dataset page: https://huggingface.co/datasets/KrononosFE/kronos-2026-2-7-cbf-safety-clamp.T145__KRONOS-8B-V7algebra-sft-1mfinMME_prockronos-r3b-dpo-pairs
Kronos R3 Branch B — DPO preference pairs
11 {prompt, chosen, rejected, question_id} pairs sampled from Qwen3-Coder-480B-A35B-Instruct (temperature=0.7, 4 candidates per problem) on LCB-medium-stdin problems, graded via _grade_public against public test cases.
chosen: shortest-passing teacher candidate
rejected: first-failing teacher candidate
filter: keep only problems where ≥1 candidate passes AND ≥1 fails
Yield notes (cycle-114/115)
30 problems sampled → 11 valid… See the full description on the dataset page: https://huggingface.co/datasets/jaivial/kronos-r3b-dpo-pairs.T145__KRONOS-8B-V3T145__KRONOS-8B-V3-details
Dataset Card for Evaluation run of T145/KRONOS-8B-V3
Dataset automatically created during the evaluation run of model T145/KRONOS-8B-V3
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/T145__KRONOS-8B-V3-details.T145__KRONOS-8B-V4-details
Dataset Card for Evaluation run of T145/KRONOS-8B-V4
Dataset automatically created during the evaluation run of model T145/KRONOS-8B-V4
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/T145__KRONOS-8B-V4-details.T145__KRONOS-8B-V8-details
Dataset Card for Evaluation run of T145/KRONOS-8B-V8
Dataset automatically created during the evaluation run of model T145/KRONOS-8B-V8
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/T145__KRONOS-8B-V8-details.T145__KRONOS-8B-V9-details
Dataset Card for Evaluation run of T145/KRONOS-8B-V9
Dataset automatically created during the evaluation run of model T145/KRONOS-8B-V9
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/T145__KRONOS-8B-V9-details.T145__KRONOS-8B-V1-P2-details
Dataset Card for Evaluation run of T145/KRONOS-8B-V1-P2
Dataset automatically created during the evaluation run of model T145/KRONOS-8B-V1-P2
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/T145__KRONOS-8B-V1-P2-details.T145__KRONOS-8B-V2-details
Dataset Card for Evaluation run of T145/KRONOS-8B-V2
Dataset automatically created during the evaluation run of model T145/KRONOS-8B-V2
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/T145__KRONOS-8B-V2-details.T145__KRONOS-8B-V5-details
Dataset Card for Evaluation run of T145/KRONOS-8B-V5
Dataset automatically created during the evaluation run of model T145/KRONOS-8B-V5
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/T145__KRONOS-8B-V5-details.T145__KRONOS-8B-V7-details
Dataset Card for Evaluation run of T145/KRONOS-8B-V7
Dataset automatically created during the evaluation run of model T145/KRONOS-8B-V7
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/T145__KRONOS-8B-V7-details.T145__KRONOS-8B-V2T145__KRONOS-8B-V6T145__KRONOS-8B-V1-P3-details
Dataset Card for Evaluation run of T145/KRONOS-8B-V1-P3
Dataset automatically created during the evaluation run of model T145/KRONOS-8B-V1-P3
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/T145__KRONOS-8B-V1-P3-details.kronos-distill-seed