MaYiding commited on
Commit
b8eb027
·
1 Parent(s): b191113

Update README

Browse files
Files changed (2) hide show
  1. README-ZH.md +1 -11
  2. README.md +0 -10
README-ZH.md CHANGED
@@ -414,19 +414,9 @@ CSV 路径完全绕过 `dataset_metadata`。要把行与提示模板配对,要
414
 
415
  ### 9.2 不适用场景
416
 
417
- * **训练数据。** 把这些行纳入任何训练、微调或 RLHF 数据集,都会污染对所训模型的下游预测评估。该数据集仅用于评估。
418
  * **长时程预测。** 所有 resolution 落在一个月窗口(2026-03-12 → 2026-04-14);该集合不代表跨季度或跨年度预测。
419
  * **开放生成。** 每题都有封闭答案集,因此并非生成基准。
420
-
421
- ### 9.3 已知局限与偏置
422
-
423
- * **样本量。** 80 行偏小。准确率或 Brier 的置信区间宽;汇报点估计时同时给出区间,并在同一集合上比较模型时使用配对检验。
424
- * **题材偏置。** 题目集中在金融与宏观指标、体育赛事、奖项(Oscars、NBA、UEFA 等)以及美国为主的政治与地缘政治事件,反映上游 HuggingFace 题集的题目结构,并非全球代表性样本。
425
- * **仅英文。** 所有 `event` 与 `options` 字符串均为英文。
426
- * **仅日期级 resolution。** `end_time` 是日期而非时间戳,且数据集不带时区列。需要更细粒度 admissibility 时,将每条 resolution 视为覆盖整个 GMT+8 日历日。
427
- * **Provider 侧残留泄漏。** 若 LLM 已摄入上游 HuggingFace 数据集,或训练截止越过 resolution 窗口,则可凭参数化记忆恢复 ground truth。数据集本身无法解决这一点,而是依赖 harness 强制 admissibility($\kappa_M$)。
428
- * **移动中的 label space 的快照。** 少数题目("none of the above"、"all of the above")与 multi-select 评分有微妙的交互;curation pass 已修复了一例 S&P 500,但未来版本对类似题目的约定可能调整。如需跨发布的字节稳定行为,请锁定到 schema version。
429
-
430
  ---
431
 
432
  ## 10. License
 
414
 
415
  ### 9.2 不适用场景
416
 
 
417
  * **长时程预测。** 所有 resolution 落在一个月窗口(2026-03-12 → 2026-04-14);该集合不代表跨季度或跨年度预测。
418
  * **开放生成。** 每题都有封闭答案集,因此并非生成基准。
419
+
 
 
 
 
 
 
 
 
 
420
  ---
421
 
422
  ## 10. License
README.md CHANGED
@@ -436,19 +436,9 @@ Without the OracleProto harness in place, treat the resulting numbers as upper b
436
 
437
  ### 9.2 Out-of-scope uses
438
 
439
- * **Training data.** Including the rows in any training, fine-tuning, or RLHF dataset contaminates downstream forecasting evaluations of the trained model. The dataset is evaluation-only.
440
  * **Long-horizon forecasting.** All resolutions land in a one-month window (2026-03-12 → 2026-04-14); the set does not represent multi-quarter or multi-year forecasting.
441
  * **Open-ended generation.** Every question has a closed answer set, so this is not a generation benchmark.
442
 
443
- ### 9.3 Known limitations and biases
444
-
445
- * **Sample size.** 80 rows is a small sample. Confidence intervals on accuracy or Brier are wide; report them alongside point estimates and use paired tests when comparing models on the same set.
446
- * **Topical skew.** Questions are concentrated in finance and macro indicators, sports events, awards (Oscars, NBA, UEFA, etc.), and US-centric political or geopolitical events, reflecting the upstream HuggingFace question mix. They are not a globally representative sample.
447
- * **English-only.** All `event` and `options` strings are in English.
448
- * **Date-only resolution.** `end_time` is a date, not a timestamp, and the dataset does not include a timezone column. If finer-grained admissibility is needed, treat each resolution as covering the whole GMT+8 calendar day.
449
- * **Provider-side residual leakage.** Any LLM that has ingested the upstream HuggingFace dataset, or that was trained past the resolution window, can recover ground truths from parametric memory. The dataset cannot address this on its own; it relies on the harness to enforce admissibility ($\kappa_M$).
450
- * **Snapshot of a moving label space.** A few questions ("none of the above", "all of the above") have subtle interactions with multi-select scoring; the curation pass fixed the one S&P 500 case, but the convention for similar questions in future revisions may shift. Pin to the schema version if byte-stable behaviour across releases is required.
451
-
452
  ---
453
 
454
  ## 10. License
 
436
 
437
  ### 9.2 Out-of-scope uses
438
 
 
439
  * **Long-horizon forecasting.** All resolutions land in a one-month window (2026-03-12 → 2026-04-14); the set does not represent multi-quarter or multi-year forecasting.
440
  * **Open-ended generation.** Every question has a closed answer set, so this is not a generation benchmark.
441
 
 
 
 
 
 
 
 
 
 
442
  ---
443
 
444
  ## 10. License