Initial v0.1 tutorial dataset

LICENSE

@@ -0,0 +1,19 @@
+MIT License
+
+Copyright (c) 2026 AuraOne
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this synthetic tutorial dataset and associated documentation files (the
+"Data"), to deal in the Data without restriction, including without limitation
+the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+sell copies of the Data, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Data.
+
+THE DATA IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
+INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
+PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
+COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
+IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+CONNECTION WITH THE DATA OR THE USE OR OTHER DEALINGS IN THE DATA.

README.md

@@ -0,0 +1,33 @@
+# EvalKit Tutorial v0.1 Dataset Card
+
+This dataset is synthetic tutorial data for exercising local EvalKit workflows. It is not expert-authored, not human-validated, not benchmark-grade, and not a safety or model-quality certification.
+
+## Intended Use
+
+Use these small fixtures to run `evalkit validate-rubric`, `evalkit lint-rubric`, `evalkit score`, report generation, and adapter mapping examples without an AuraOne account or API key.
+
+## Out Of Scope Use
+
+Do not use this data to compare models, publish benchmark claims, certify safety, or represent customer reviewer performance.
+
+## Files
+
+- `rubric.jsonl`: synthetic code-review tutorial rubric.
+- `model_outputs.jsonl`: synthetic tutorial model outputs.
+- `labels.jsonl`: synthetic tutorial labels.
+- `expected_scores.json`: deterministic expected score summary.
+- `LICENSE`: MIT-compatible fixture license.
+
+## Synthetic Generation Process
+
+Rows were hand-written as low-risk code-review-style tutorial examples. They intentionally avoid medical, legal, safety-critical, customer, reviewer identity, and production data.
+
+## Limitations
+
+- Small sample count.
+- Not validated by domain experts.
+- Public tutorial data can leak into prompts and should not be used as a hidden eval.
+
+## Citation
+
+Cite AuraOne EvalKit documentation when reusing the template or adapting the schema.

expected_scores.json

@@ -0,0 +1,6 @@
+{
+  "average_score": 0.645833,
+  "pass_rate": 0.666667,
+  "scored_outputs": 3,
+  "data_status": "synthetic tutorial data; not expert-authored, human-validated, or benchmark-grade"
+}

labels.jsonl

@@ -0,0 +1,12 @@
+{"output_id":"out-001","criterion_id":"code_review.correctness","score":1.0,"label_source":"synthetic_tutorial"}
+{"output_id":"out-001","criterion_id":"code_review.specificity","score":0.75,"label_source":"synthetic_tutorial"}
+{"output_id":"out-001","criterion_id":"code_review.actionability","score":0.75,"label_source":"synthetic_tutorial"}
+{"output_id":"out-001","criterion_id":"code_review.tone","score":0.75,"label_source":"synthetic_tutorial"}
+{"output_id":"out-002","criterion_id":"code_review.correctness","score":0.25,"label_source":"synthetic_tutorial"}
+{"output_id":"out-002","criterion_id":"code_review.specificity","score":0.0,"label_source":"synthetic_tutorial"}
+{"output_id":"out-002","criterion_id":"code_review.actionability","score":0.25,"label_source":"synthetic_tutorial"}
+{"output_id":"out-002","criterion_id":"code_review.tone","score":1.0,"label_source":"synthetic_tutorial"}
+{"output_id":"out-003","criterion_id":"code_review.correctness","score":0.8,"label_source":"synthetic_tutorial"}
+{"output_id":"out-003","criterion_id":"code_review.specificity","score":0.8,"label_source":"synthetic_tutorial"}
+{"output_id":"out-003","criterion_id":"code_review.actionability","score":0.8,"label_source":"synthetic_tutorial"}
+{"output_id":"out-003","criterion_id":"code_review.tone","score":0.8,"label_source":"synthetic_tutorial"}

model_outputs.jsonl

@@ -0,0 +1,3 @@
+{"output_id":"out-001","prompt":"Review this synthetic parser change.","output":"The change handles empty input and cites parser.py, but it logs a token in debug mode and needs a regression test.","model":"tutorial-model-a","metadata":{"benchmark":false,"dataset":"synthetic tutorial"}}
+{"output_id":"out-002","prompt":"Review this synthetic API change.","output":"Looks fine overall. Maybe rename a variable.","model":"tutorial-model-b","metadata":{"benchmark":false,"dataset":"synthetic tutorial"}}
+{"output_id":"out-003","prompt":"Review this synthetic migration.","output":"The review names the compatibility risk, cites migration.ts, and asks for one targeted test in a professional tone.","model":"tutorial-model-c","metadata":{"benchmark":false,"dataset":"synthetic tutorial"}}

rubric.jsonl

@@ -0,0 +1,4 @@
+{"criterion_id":"code_review.correctness","domain":"code_review_tutorial","task_type":"code_review","criterion":"Identify whether the review catches concrete correctness risks in the supplied synthetic code change.","weight":0.35,"severity":"error","scoring_type":"scale_0_1","examples":[{"positive":"Names an input boundary case that can fail.","negative":"Only says the implementation looks fine."}],"edge_cases":["empty inputs","duplicate identifiers"],"disagreement_risk":{"level":"medium","notes":"Reviewers may disagree on partial correctness issues."},"data_source":"synthetic_tutorial","tags":["synthetic","tutorial","correctness"],"version":"0.1.0","notes":"Synthetic tutorial criterion; not expert-authored.","score_levels":{"0":"misses correctness risk","0.5":"mentions a vague risk","1":"identifies a concrete correctness risk"}}
+{"criterion_id":"code_review.specificity","domain":"code_review_tutorial","task_type":"code_review","criterion":"Cite specific files, functions, tests, or output evidence that would let a reader verify the review claim.","weight":0.25,"severity":"warning","scoring_type":"scale_0_1","examples":[{"positive":"References parser.py and the missing empty-list test.","negative":"Makes a general claim without evidence."}],"edge_cases":["documentation-only changes","generated files"],"disagreement_risk":{"level":"low","notes":"Evidence references are visible in the review text."},"data_source":"synthetic_tutorial","tags":["synthetic","tutorial","evidence"],"version":"0.1.0","notes":"Synthetic tutorial criterion; not expert-authored.","score_levels":{"0":"no evidence","0.5":"partial evidence","1":"specific verifiable evidence"}}
+{"criterion_id":"code_review.actionability","domain":"code_review_tutorial","task_type":"code_review","criterion":"Separate required fixes from optional refactors and provide an action the implementer can take next.","weight":0.25,"severity":"warning","scoring_type":"scale_0_1","examples":[{"positive":"Says add a regression test before merging.","negative":"Requests a broad rewrite without a concrete next step."}],"edge_cases":["style-only comments","large refactors"],"disagreement_risk":{"level":"medium","notes":"Actionability can vary by team convention."},"data_source":"synthetic_tutorial","tags":["synthetic","tutorial","actionability"],"version":"0.1.0","notes":"Synthetic tutorial criterion; not expert-authored.","score_levels":{"0":"not actionable","0.5":"partly actionable","1":"clear next action"}}
+{"criterion_id":"code_review.tone","domain":"code_review_tutorial","task_type":"code_review","criterion":"Use direct professional language that explains risk without insulting the author or overstating certainty.","weight":0.15,"severity":"info","scoring_type":"scale_0_1","examples":[{"positive":"States the risk and suggested fix directly.","negative":"Uses dismissive language or certainty not supported by evidence."}],"edge_cases":["urgent production bug","security-sensitive review"],"disagreement_risk":{"level":"low","notes":"Tone issues are visible in the review text."},"data_source":"synthetic_tutorial","tags":["synthetic","tutorial","communication"],"version":"0.1.0","notes":"Synthetic tutorial criterion; not expert-authored.","score_levels":{"0":"unprofessional or overstated","0.5":"mixed tone","1":"direct and professional"}}