Datasets:
|
Download README.md from lfqian/Annulus-Economic-Benchmark: direct link, hf CLI and curl.
- Browser
- Download file 25.9 kB
-
https://huggingface.co/datasets/lfqian/Annulus-Economic-Benchmark/resolve/main/README.md
- Command line
-
hf download hf://datasets/lfqian/Annulus-Economic-Benchmark/README.md
-
curl -L -o README.md https://huggingface.co/datasets/lfqian/Annulus-Economic-Benchmark/resolve/main/README.md
25.9 kB
| language: | |
| - en | |
| license: other | |
| license_name: source-dependent | |
| license_link: https://huggingface.co/datasets/lfqian/Annulus-Economic-Benchmark/blob/main/SOURCE_NOTICES.md | |
| task_categories: | |
| - question-answering | |
| - text-generation | |
| pretty_name: Annulus Economic Benchmark | |
| size_categories: | |
| - 10K<n<100K | |
| tags: | |
| - finance | |
| - economics | |
| - forecasting | |
| - temporal-reasoning | |
| - knowledge-boundary | |
| - benchmark | |
| configs: | |
| - config_name: task1_1990 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_1990.parquet | |
| default: true | |
| - config_name: task1_1991 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_1991.parquet | |
| - config_name: task1_1992 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_1992.parquet | |
| - config_name: task1_1993 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_1993.parquet | |
| - config_name: task1_1994 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_1994.parquet | |
| - config_name: task1_1995 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_1995.parquet | |
| - config_name: task1_1996 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_1996.parquet | |
| - config_name: task1_1997 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_1997.parquet | |
| - config_name: task1_1998 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_1998.parquet | |
| - config_name: task1_1999 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_1999.parquet | |
| - config_name: task1_2000 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2000.parquet | |
| - config_name: task1_2001 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2001.parquet | |
| - config_name: task1_2002 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2002.parquet | |
| - config_name: task1_2003 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2003.parquet | |
| - config_name: task1_2004 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2004.parquet | |
| - config_name: task1_2005 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2005.parquet | |
| - config_name: task1_2006 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2006.parquet | |
| - config_name: task1_2007 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2007.parquet | |
| - config_name: task1_2008 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2008.parquet | |
| - config_name: task1_2009 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2009.parquet | |
| - config_name: task1_2010 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2010.parquet | |
| - config_name: task1_2011 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2011.parquet | |
| - config_name: task1_2012 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2012.parquet | |
| - config_name: task1_2013 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2013.parquet | |
| - config_name: task1_2014 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2014.parquet | |
| - config_name: task1_2015 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2015.parquet | |
| - config_name: task1_2016 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2016.parquet | |
| - config_name: task1_2017 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2017.parquet | |
| - config_name: task1_2018 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2018.parquet | |
| - config_name: task1_2019 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2019.parquet | |
| - config_name: task1_2020 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2020.parquet | |
| - config_name: task1_2021 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2021.parquet | |
| - config_name: task1_2022 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2022.parquet | |
| - config_name: task1_2023 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2023.parquet | |
| - config_name: task1_2024 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2024.parquet | |
| - config_name: task1_2025 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2025.parquet | |
| - config_name: task1_2026 | |
| data_files: | |
| - split: test | |
| path: viewer/task1/year_2026.parquet | |
| - config_name: task2_1990 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_1990.parquet | |
| - config_name: task2_1991 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_1991.parquet | |
| - config_name: task2_1992 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_1992.parquet | |
| - config_name: task2_1993 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_1993.parquet | |
| - config_name: task2_1994 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_1994.parquet | |
| - config_name: task2_1995 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_1995.parquet | |
| - config_name: task2_1996 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_1996.parquet | |
| - config_name: task2_1997 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_1997.parquet | |
| - config_name: task2_1998 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_1998.parquet | |
| - config_name: task2_1999 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_1999.parquet | |
| - config_name: task2_2000 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2000.parquet | |
| - config_name: task2_2001 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2001.parquet | |
| - config_name: task2_2002 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2002.parquet | |
| - config_name: task2_2003 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2003.parquet | |
| - config_name: task2_2004 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2004.parquet | |
| - config_name: task2_2005 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2005.parquet | |
| - config_name: task2_2006 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2006.parquet | |
| - config_name: task2_2007 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2007.parquet | |
| - config_name: task2_2008 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2008.parquet | |
| - config_name: task2_2009 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2009.parquet | |
| - config_name: task2_2010 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2010.parquet | |
| - config_name: task2_2011 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2011.parquet | |
| - config_name: task2_2012 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2012.parquet | |
| - config_name: task2_2013 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2013.parquet | |
| - config_name: task2_2014 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2014.parquet | |
| - config_name: task2_2015 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2015.parquet | |
| - config_name: task2_2016 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2016.parquet | |
| - config_name: task2_2017 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2017.parquet | |
| - config_name: task2_2018 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2018.parquet | |
| - config_name: task2_2019 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2019.parquet | |
| - config_name: task2_2020 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2020.parquet | |
| - config_name: task2_2021 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2021.parquet | |
| - config_name: task2_2022 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2022.parquet | |
| - config_name: task2_2023 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2023.parquet | |
| - config_name: task2_2024 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2024.parquet | |
| - config_name: task2_2025 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2025.parquet | |
| - config_name: task2_2026 | |
| data_files: | |
| - split: test | |
| path: viewer/task2/year_2026.parquet | |
| - config_name: task3_2023 | |
| data_files: | |
| - split: test | |
| path: viewer/task3/year_2023.parquet | |
| - config_name: task3_2024 | |
| data_files: | |
| - split: test | |
| path: viewer/task3/year_2024.parquet | |
| - config_name: task3_2025 | |
| data_files: | |
| - split: test | |
| path: viewer/task3/year_2025.parquet | |
| - config_name: task3_2026 | |
| data_files: | |
| - split: test | |
| path: viewer/task3/year_2026.parquet | |
| - config_name: findendro_recipe | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_*.parquet | |
| - config_name: findendro_facts | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/facts/year_*.parquet | |
| - config_name: findendro_recipe_1999 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_1999.parquet | |
| - config_name: findendro_recipe_2000 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2000.parquet | |
| - config_name: findendro_recipe_2001 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2001.parquet | |
| - config_name: findendro_recipe_2002 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2002.parquet | |
| - config_name: findendro_recipe_2003 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2003.parquet | |
| - config_name: findendro_recipe_2004 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2004.parquet | |
| - config_name: findendro_recipe_2005 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2005.parquet | |
| - config_name: findendro_recipe_2006 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2006.parquet | |
| - config_name: findendro_recipe_2007 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2007.parquet | |
| - config_name: findendro_recipe_2008 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2008.parquet | |
| - config_name: findendro_recipe_2009 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2009.parquet | |
| - config_name: findendro_recipe_2010 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2010.parquet | |
| - config_name: findendro_recipe_2011 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2011.parquet | |
| - config_name: findendro_recipe_2012 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2012.parquet | |
| - config_name: findendro_recipe_2013 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2013.parquet | |
| - config_name: findendro_recipe_2014 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2014.parquet | |
| - config_name: findendro_recipe_2015 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2015.parquet | |
| - config_name: findendro_recipe_2016 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2016.parquet | |
| - config_name: findendro_recipe_2017 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2017.parquet | |
| - config_name: findendro_recipe_2018 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2018.parquet | |
| - config_name: findendro_recipe_2019 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2019.parquet | |
| - config_name: findendro_recipe_2020 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2020.parquet | |
| - config_name: findendro_recipe_2021 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2021.parquet | |
| - config_name: findendro_recipe_2022 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2022.parquet | |
| - config_name: findendro_recipe_2023 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2023.parquet | |
| - config_name: findendro_recipe_2024 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2024.parquet | |
| - config_name: findendro_recipe_2025 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2025.parquet | |
| - config_name: findendro_recipe_2026 | |
| data_files: | |
| - split: test | |
| path: findendro_recipe_replica/20261024/viewer/prompts/year_2026.parquet | |
| # Annulus Economic Benchmark | |
| Annulus Economic Benchmark contains three tasks built from economic statistical series: self-reported knowledge boundaries, historical value recall, and retrospective monthly forecasts. All model-facing prompts are in English. This release contains 14,040 fact records and 50,220 questions. | |
| ## Coverage | |
| | Component | Target periods | Facts per period | Task 1 per period | Task 2 per period | Task 3 per period | | |
| | --- | --- | ---: | ---: | ---: | ---: | | |
| | Annual | 1990–2022, inclusive | 180 | 360 | 180 | — | | |
| | Monthly | January 2023–September 2026, inclusive | 180 | 360 | 180 | 180 | | |
| The annual component contains 5,940 facts and 17,820 questions. The monthly component contains 8,100 facts and 32,400 questions. Across both components, Task 1 has 28,080 questions, Task 2 has 14,040, and Task 3 has 8,100. The annual and monthly panels differ; within a monthly target period, the three tasks use the same country–indicator objects. | |
| ## Tasks | |
| Task 1 examines whether a model claims to know an outcome or treats it as a forecast. Task 2 checks recall of the recorded value. Task 3 elicits probabilities for the next month's direction of change. | |
| **Task 1 — Self-reported knowledge boundary.** Each fact produces a recall question and a prediction question. The choices distinguish factual answers, inability to recall an outcome outside the model’s knowledge, and prediction rather than claimed prior knowledge. Gold labels are conditional on the `inside` or `outside` boundary condition. A reference month, year, or advertised model cutoff does not by itself establish whether a particular fact is inside a model’s knowledge. Users must specify their boundary-assignment protocol when reporting Task 1 scores. | |
| **Task 2 — Historical value recall.** Each question asks for one economic indicator's value for a specified annual or monthly reference period, with four numerical answer choices. The model selects the recorded value, rounded to two decimal places as stated in the question. Task 2 contains 14,040 ordinary factual-recall questions: 180 per annual target year and 180 per target month from 2023 onward. The reference period does not establish the value's first-publication date or the complete information set available at that time. | |
| **Task 3 — Monthly forecast.** The forecast date is the end of the preceding month. The model predicts whether the target month’s statistical value will increase, remain unchanged, or decrease relative to the preceding month. It must return a probability for every labelled choice, summing to one, and a short `reason`. These are retrospective forecast questions with recorded outcomes, not a claim that the historical information available at each forecast date has been fully reconstructed. | |
| ## Sources and construction | |
| The data are constructed from frozen statistical-table downloads from the Bank for International Settlements, World Bank WDI and GEM, OECD, and national statistical or central-bank publications. The release includes 50 verified supplemental observations, with source provenance and any quotation conversions retained in the evidence. Supplements include explicitly identified attributed sources where an official original was inaccessible. Source values are used to construct numerical choices, factual-recall answers, and forecast outcomes. | |
| Five topics were replaced only in September 2026 to use available published outcomes while preserving the regional counts: Japan CPI → Japan long-term government bond yield; UK CPI → UK long-term government bond yield; South Africa CPI → Kenya policy rate; Morocco CPI → Nigeria policy rate; and Algeria CPI → Ghana policy rate. These substitutions are recorded in the release metadata and should be considered when comparing September with earlier months. | |
| ## Two-column evaluation table | |
| Select the task and target year together under **Subset**, for example `task1_1990`, `task2_1990` or `task3_2026`. **Split** is always `test`. Every displayed row has exactly two top-level columns: | |
| | Column | Content | Use | | |
| | --- | --- | --- | | |
| | `input` | The complete prompt as a plain text string, including instructions, the question and labelled options. | Pass this string directly to the model. | | |
| | `gold` | The reference answer for that same row. | Read only for offline scoring. | | |
| Task 1's `gold` is an object with the existing `inside` and `outside` option letters. Select the condition according to your boundary-evaluation protocol; the release does not guess a model-specific knowledge boundary from a date. Task 2's `gold` is the option letter for the correct numerical value. Task 3's `gold` is the observed outcome's option letter; its input still requests a probability distribution and a reason. Task 3's mapping from option letters to directions is retained in each input, so a particular letter must not be assumed always to mean an increase or decrease. | |
| Task 1 and Task 2 each have 37 task/year configurations, covering 1990–2026. Task 3 has `task3_2023`, `task3_2024`, `task3_2025` and `task3_2026`. There are 78 configurations in total, each with one `test` split. Years are configurations rather than split names so that the Dataset Viewer does not encounter dozens of nonstandard splits. The month remains written in each monthly question. | |
| ## Validation and limitations | |
| Numerical values, derived answers, and export consistency have been checked against the frozen source records and documented supplements. The release has not undergone a complete item-by-item semantic review or a full LLM rewriting pass; the questions are generated deterministically from the statistical records. | |
| The source snapshots may contain revisions and are not historical real-time vintages. First-publication dates have not been comprehensively verified. Reference-period labels and evaluation dates must not be treated as public-release dates: annual series may follow fiscal-year reporting conventions, and a December 31 grouping date does not establish a release date. Monthly interest-rate observations are not all month-end rates; each series retains its stated measurement basis. Supplemented exchange-rate observations retain their quotation basis, transformations, and effective precision. | |
| The dataset provides prompts, outcomes, and provenance for evaluation. This release does not include model-evaluation results or establish a model’s actual knowledge state from calendar dates alone. | |
| ## Loading and running | |
| ```python | |
| from datasets import load_dataset | |
| ds = load_dataset( | |
| "lfqian/Annulus-Economic-Benchmark", | |
| "task1_1990", | |
| split="test", | |
| token=False, | |
| ) | |
| row = ds[0] | |
| prompt = row["input"] # Plain text: pass this directly to your model. | |
| gold = row["gold"] # Keep separate for offline scoring. | |
| ``` | |
| The prompts are in English. Task 2 now uses four-option value-recall questions built from the existing fact records. The two-column projection contains all 50,220 benchmark questions. | |
| ## Auxiliary files | |
| The full annotation tables under `data/{task}/year_YYYY.parquet` retain IDs, metadata and source evidence. The corresponding `viewer/{task}/year_YYYY.parquet` tables contain only `input` and `gold`, in exactly the same row order. | |
| The optional `evaluation/{task}/{year}/input.jsonl` and `gold.jsonl` pairs retain IDs for batch runners that join results by ID. Task 2 IDs use `{fact_id}/state`. The `data/facts` and `model_inputs` folders retain the underlying fact records and prompt-only copies. These auxiliary files are not additional tasks, Dataset Viewer configurations, or benchmark questions. Earlier Task 2 versions remain available in the Hugging Face repository's version history. | |
| `metadata/release.json` records counts, task/year configurations, regional quotas, topic substitutions, validation status and the hashes of numerical audit reports. `metadata/alignments.jsonl` connects tasks by object and target period. `metadata/evaluation_pairs.json` lists the optional JSONL pairs. `metadata/panels.json` describes the sources, and `metadata/checksums.json` binds the public files. See [source notices](SOURCE_NOTICES.md) for attribution and reuse status. | |
| ## Language | |
| All model-facing instructions, questions, and answer choices are in English. The underlying source values and provenance are preserved. Quoted source excerpts in the auxiliary evidence retain their original language. | |
| ## Option and prompt revision (2026-10-22) | |
| Task 1 uses complete knowledge-state statements of comparable length. Numerical distractors use randomized truth ranks and unequal random distances, rather than a fixed arithmetic progression. All three tasks shuffle options independently per question; question order is shuffled within each task/year. Gold letters and all input/gold exports are updated together. Task 3 specifies the JSON fields without sample probability values and requests a brief reason based on information at the forecast origin. The seed is recorded in metadata/release.json for reproduction. Source observations, answers, periods and question counts are unchanged. Earlier gold letters must not be reused. | |
| ## Monthly coverage | |
| Each month from January 2023 through September 2026 contains 180 distinct country–indicator objects: 180 fact records, 360 Task 1 questions (180 Recall and 180 Predict), 180 Task 2 questions and 180 Task 3 questions. The additional panel uses complete BIS month-end exchange rates and policy rates and OECD monthly market interest rates. Every added item retains its source snapshot, SHA-256, CSV row, unit and reference period. Shared euro exchange-rate records are excluded; identical benchmark-rate paths are counted once among added candidates. No outcome-direction balancing is applied. All answer labels follow the independently shuffled displayed options. | |
| Monthly regional counts are Africa: 22, Oceania: 8, South America: 16, Asia: 53, Europe: 64, North America: 17. This allocation prioritizes complete official source coverage over exact regional proportionality. | |
| <!-- findendro-recipe-20261024:start --> | |
| ## FinDendro recipe replica (2026-10-24) | |
| This independently reconstructed version follows the paper recipe. It is **not the original author sample or original model results**. | |
| | Reference year | Facts per year | Prompts per year | | |
| |---|---:|---:| | |
| | 1999–2020 | 48 | 240 | | |
| | 2021–2024 | 204 | 1,020 | | |
| | 2025 | 204 | 1,020 | | |
| | 2026 (January–September) | 153 | 765 | | |
| | Total | 2,229 | 11,145 | | |
| The sample includes 1,356 macroeconomic facts, 540 stock-price facts and 333 Federal Reserve policy facts. The real observation pool contains 20,130 records. Each fact has five conditions: value, recall, predict, recall_asof, predict_asof. | |
| Use the **findendro_recipe** config for all 11,145 browsable prompts, **findendro_facts** for the 2,229 sampled facts, or **findendro_recipe_YYYY** for one reference year. Existing configurations and the repository's default remain available. | |
| ```python | |
| from datasets import load_dataset | |
| prompts = load_dataset("lfqian/Annulus-Economic-Benchmark", "findendro_recipe", split="test") | |
| year_2025 = load_dataset("lfqian/Annulus-Economic-Benchmark", "findendro_recipe_2025", split="test") | |
| ``` | |
| - [Version directory](findendro_recipe_replica/20261024) | |
| - [Full archive with raw snapshots and construction code](findendro_recipe_replica/20261024/findendro-replica-20261024.zip) | |
| - [Per-year alignment report](findendro_recipe_replica/20261024/metadata/year_alignment.md) | |
| - [Complete five-condition example](findendro_recipe_replica/20261024/example_prompts.txt) | |
| - [Construction configuration](findendro_recipe_replica/20261024/metadata/config.json) | |
| - [Same-month substitute sources](findendro_recipe_replica/20261024/metadata/supplemental_manifest.json) | |
| - [Substitute observation checks](findendro_recipe_replica/20261024/metadata/supplement_check.json) | |
| For model evaluation use `findendro_recipe_replica/20261024/evaluation/YYYY/input.jsonl`; factual answer mappings are separately stored in `gold.jsonl`. Viewer Parquet files include answer keys for browsing, and should not be passed wholesale to models. Knowledge claims are decoded separately from factual correctness; no ground-truth internal knowledge label is invented. | |
| Years follow the fact's reference month. January as-of prompts intentionally specify December 31 of the preceding year. All question/fact/gold joins and yearly partitions passed validation. | |
| **October 2025 uses three real same-month substitute observations:** one-, five-, and thirty-year Treasury yields (`GS1`, `GS5`, `GS30`). They occupy the three missing macro sample slots under their own series IDs. The original October `CPIAUCNS`, `CPILFENS` and `UNRATENSA` observations **remain missing**; no CPI or unemployment values are fabricated or imputed. The original missingness is documented by [BLS CPI](https://www.bls.gov/cpi/additional-resources/2025-federal-government-shutdown-impact-cpi.htm) and [BLS CPS](https://www.bls.gov/cps/methods/2025-federal-government-shutdown-impact-cps.htm). This explicit substitution extends the paper recipe and is documented in `metadata/supplemental_manifest.json` and `metadata/supplement_check.json`. | |
| The persistence-controlled subset contains 1,335 facts and has an aggregate random baseline of approximately 0.2708. All sources, sampling choices, option generation rules and missingness are documented in the version directory. These are real retrieved historical snapshots; **original first-release vintages have not been independently verified**. Data reuse remains source-dependent. | |
| <!-- findendro-recipe-20261024:end --> | |