Update README.md

README.md

@@ -8,16 +8,21 @@ size_categories:
 - 1K<n<10K
 ---
 # KOCOH (KOrean COntext-dependent Hate speech) Dataset
-* Paper: [KOCOH: A Dataset for Detecting Context-Dependent Hate Speech](https://www.kci.go.kr/kciportal/ci/sereArticleSearch/ciSereArtiView.kci?sereArticleSearchBean.artiId=ART003173761)
+* Paper: [KOCOH: Korean Context-Dependent Hate Speech Dataset](https://doi.org/10.63317/5pnu2jn6awun)
 * Authors: [Park Eunah](https://github.com/eparkatgithub), [Song Sanghoun](http://corpus.mireene.com/)
 * Contact: dmsdk1993@korea.ac.kr
 * [GitHub](https://github.com/eparkatgithub/KOCOH)😺
 
-This is **Ko**rean **co**ntext-dependent **h**ate speech dataset, KOCOH.
+This is **Ko**rean **co**ntext-dependent **h**ate speech dataset, KOCOH.<br>
+[한국어 설명](https://github.com/eparkatgithub/KOCOH/blob/main/README_ko.md)
+
+## Histroy
+* 2025/10/24: Update to Version 2
+* 2025/01/31: First public release of the dataset
 
 ## Main concept
 * Hate speech
-  * Verbal or non-verbal expressions that propagate or promote prejudice or discrimination against minorities (groups with common identity who have relatively diminished political and social power), or denigrate, insult, or threaten individuals or groups based on their attributes as minorities, or incites discrimination, hostility, or violence against them (Hong, 2018)
+  * Verbal or non-verbal expressions that, based on prejudice or discrimination against socially marginalized or vulnerable groups, denigrate, disparage, or insult such groups or individuals belonging to them, or incite discrimination or violence against them.
 * Context-dependent hate speech
   * Hate speech interpreted as hateful through contextual factors rather than explicit content alone
   * In other words, without context, the statement itself may appear neutral or even positive in interpretation
@@ -28,12 +33,13 @@ This is **Ko**rean **co**ntext-dependent **h**ate speech dataset, KOCOH.
     |직원에게 높은 수준의 복지를 제공한 기업의 이름이 공개되었다.<br>*A list of companies that provide high-level welfare benefits to employees was released.*|믿을 만한 그룹이라는 얘기네<br>*This means they're a trustworthy group.*|0|
 
 ## Data Description
-* Source: [Dcinside Real-time Best Gallery](https://gall.dcinside.com/board/lists/?id=dcbest)
-* Period: 2024/06/02-23
+* Source
+  * [Dcinside Real-time Best Gallery](https://gall.dcinside.com/board/lists/?id=dcbest) (2024/06/02-23)
+  * [FMkorea Best Board](https://www.fmkorea.com/best)(2025/01/01-21)
 * Size
   |Total|Type 1|Type 2|Type 3|
   |:---:|:---:|:---:|:---:|
-  |2,005|539|539|927|
+  |3,000|764|764|1,472|
 * Types and examples
   <table>
     <thead>
@@ -87,80 +93,89 @@ This is **Ko**rean **co**ntext-dependent **h**ate speech dataset, KOCOH.
     </thead>
     <tbody>
       <tr>
-        <th>index</th>
+        <th>Index</th>
         <td>Data index</td>
       </tr>
       <tr>
-        <th>set</th>
+        <th>Set</th>
         <td>Post index</td>
       </tr>
       <tr>
-        <th>type</th>
+        <th>Type</th>
         <td>Type number (1~3 labeling)</td>
       </tr>
       <tr>
-        <th>date</th>
-        <td>The date the DC Inside post was written</td>
+        <th>Date</th>
+        <td>The date the post was written</td>
       </tr>
       <tr>
-        <th>link</th>
-        <td>The link to the DC Inside post</td>
+        <th>Link</th>
+        <td>The link to the post</td>
       </tr>
       <tr>
-        <th>title</th>
-        <td>The title of the DC Inside post</td>
+        <th>Title</th>
+        <td>The title of the post</td>
       </tr>
       <tr>
-        <th>context</th>
-        <td>Summary of the DC Inside post content / Created context</td>
+        <th>Context</th>
+        <td>Summary of the post content / Created context</td>
       </tr>
       <tr>
-        <th>comment</th>
-        <td>Comments collected from DC Inside</td>
+        <th>Comment</th>
+        <td>Comments collected from the forum</td>
       </tr>
       <tr>
-        <th>hate speech</th>
+        <th>Hate speech</th>
         <td>Whether it is hate speech (0 or 1 labeling)</td>
       </tr>
       <tr>
-        <th>gender</th>
+        <th>Counter speech</th>
+        <td>Whether it is counter speech (0 or 1 labeling)</td>
+      </tr>
+      <tr>
+        <th>Gender</th>
         <td>Target: gender (0 or 1 labeling)</td>
       </tr>
       <tr>
-        <th>disability</th>
+        <th>Disability</th>
         <td>Target: disability (0 or 1 labeling)</td>
       </tr>
       <tr>
-        <th>race/nation</th>
-        <td>Target: race/nation (0 or 1 labeling)</td>
+        <th>Race/Nationality</th>
+        <td>Target: race/nationality (0 or 1 labeling)</td>
       </tr>
       <tr>
-        <th>region</th>
-        <td>Target: region (0 or 1 labeling)</td>
+        <th>Region (Korea)</th>
+        <td>Target: region (Korea) (0 or 1 labeling)</td>
       </tr>
       <tr>
-        <th>age</th>
+        <th>Age</th>
         <td>Target: age (0 or 1 labeling)</td>
       </tr>
       <tr>
-        <th>note</th>
-        <td>Ikiyano style (-no)</td>
+        <th>Profanity</th>
+        <td>Including profanity (0 or 1 labeling)</td>
+      </tr>
+      <tr>
+        <th>Ikiyano-style</th>
+        <td>Including <i>ikiyano</i> style (-no) (0 or 1 labeling)</td>
       </tr>
     </tbody>
   </table>
 
-## Reference
-Hong, S. (2018). [*When Words Hurt*](https://scholarworks.sookmyung.ac.kr/handle/2020.sw.sookmyung/20524). Across.
-
 ## Citation
 ```
-@article{ART003173761,
-author={박은아 and 송상헌},
-title={KOCOH: 맥락 의존적 혐오 표현 탐지를 위한 데이터 세트},
-journal={한국어학},
-issn={1226-9123},
-year={2025},
-volume={106},
-pages={251-277}
+@inproceedings{park-etal-2026-kocoh,
+  title = {KOCOH: Korean Context-Dependent Hate Speech Dataset},
+  author = {Park, Eunah and Song, Sanghoun},
+  booktitle = {Proceedings of the Fifteenth Language Resources and Evaluation Conference (LREC 2026)},
+  month = {May},
+  year = {2026},
+  pages = {4103--4114},
+  address = {Palma, Mallorca, Spain},
+  publisher = {European Language Resources Association (ELRA)},
+  editor = {Piperidis, Stelios and Bel, Núria and van den Heuvel, Henk and Ide, Nancy and Krek, Simon and Toral, Antonio},
+  doi = {10.63317/5pnu2jn6awun},
+  abstract = {We introduce the KOrean COntext-dependent Hate speech dataset (KOCOH) to evaluate large language models’ ability to detect context-dependent hate speech in Korean. KOCOH consists of 3,000 context-comment pairs collected from Korean online communities (Dcinside, FMkorea) with detailed annotations, including labels for hate speech and hate target groups. We assess the context-dependent hate speech detection capabilities of both humans and 11 state-of-the-art large language models, including GPT-5, Claude Sonnet 4, and Gemini 2.5 Flash. Our results show that humans outperform language models, with GPT-5 achieving the highest performance among the evaluated models. While humans demonstrate balanced recall and specificity, language models generally show significantly higher specificity compared to recall. The performance of both humans and models is affected by factors such as Honam-related vocabulary and sentiment polarity. This study contributes resources to Korean hate speech research and empirically demonstrates the performance gap between humans and language models. Through both quantitative and qualitative analyses, we explore the similarities and differences between humans and language models, offering insights for future developments in language models and AI ethics research. KOCOH is available at https://github.com/eparkatgithub/KOCOH.}
 }
 ```