mkd-minju commited on
Commit
f5ed806
·
verified ·
1 Parent(s): f8df427

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +20 -1
README.md CHANGED
@@ -15,7 +15,7 @@ task_categories:
15
 
16
  <p align="center">
17
  <a href="#1-introduction">Introduction</a> |
18
- <a href="#2-dataset-composition">Dataset Composition</a> |
19
  <a href="#3-data-collection--processing-methodology">Methodology</a> |
20
  <a href="#4-license-and-attribution">License</a> |
21
  <a href="#5-known-limitations">Limitations</a>
@@ -63,6 +63,25 @@ The corpus totals **300,000 examples**, split 80/20 between Korean (240,000) and
63
 
64
  This slice was deliberately restricted to human-authored data. Given how much of the rest of the Keural-v2 corpus is necessarily LLM-synthesized, the English balance slice is held to a stricter standard specifically to avoid compounding synthetic-data risk in the one category whose only job is to preserve an already-solid baseline capability.
65
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
66
  ## 3. Data Collection & Processing Methodology
67
 
68
  1. Each source was downloaded from its canonical location and exact-match deduplicated internally on its primary text field.
 
15
 
16
  <p align="center">
17
  <a href="#1-introduction">Introduction</a> |
18
+ <a href="#2-dataset-composition">Dataset Composition</a> (<a href="#23-source-revisions">Revisions</a>) |
19
  <a href="#3-data-collection--processing-methodology">Methodology</a> |
20
  <a href="#4-license-and-attribution">License</a> |
21
  <a href="#5-known-limitations">Limitations</a>
 
63
 
64
  This slice was deliberately restricted to human-authored data. Given how much of the rest of the Keural-v2 corpus is necessarily LLM-synthesized, the English balance slice is held to a stricter standard specifically to avoid compounding synthetic-data risk in the one category whose only job is to preserve an already-solid baseline capability.
65
 
66
+ ### 2.3 Source Revisions
67
+
68
+ Exact commit/revision pinned at download time (2026-08-06), for reproducibility and future re-verification:
69
+
70
+ <div align="center">
71
+
72
+ | Source | Revision |
73
+ |:---|:---|
74
+ | `mkd-chanwoo/keural-conversation-ko` | `fcb975654651d55ff1a642455d620045b6d3f741` |
75
+ | `IkJun1/korean-qa-dataset` | `e1f177a7497cf4e55e54d86101c6c522345441d2` |
76
+ | `coastral/korean-writing-style-instruct` | `9f31462314f07847534fd13f7406c330ab15d156` |
77
+ | `CarrotAI/ko-instruction-dataset` | `5c0e2c0180b50400e401dd0b296043f18fc6cb3f` |
78
+ | `beomi/KoAlpaca-RealQA` | `a7df38a0b2cc187b72b40330af81e7b9f28dd95b` |
79
+ | `databricks/databricks-dolly-15k` | `bdd27f4d94b9c1f951818a7da7fd7aeea5dbff1a` |
80
+ | `OpenAssistant/oasst2` | `179dd21fc55192153d94adb0e0ce8f69e222bf75` |
81
+ | Taskmaster-1 / 2 / 3, MultiWOZ 2.2 | Fetched directly from `github.com/google-research-datasets/Taskmaster` and `github.com/budzianowski/multiwoz` (`master` branch, 2026-08-06) — these are legacy loading-script-only HF repos with no data-bearing revision of their own, so the upstream GitHub source is the canonical reference |
82
+
83
+ </div>
84
+
85
  ## 3. Data Collection & Processing Methodology
86
 
87
  1. Each source was downloaded from its canonical location and exact-match deduplicated internally on its primary text field.