Update README.md

README.md

@@ -21,23 +21,23 @@ In such a pre-training paradigm,
 
 
 ## Model Description
-We release all models introduced in our [paper](), covering 13 different application scenarios. Each model contains 11 billion parameters.
+We release all models introduced in our [paper](https://arxiv.org/pdf/2206.11147.pdf), covering 13 different application scenarios. Each model contains 11 billion parameters.
 
 | Model      | Description | Recommended Application
 | ----------- | ----------- |----------- |
-| **rst-all-11b**   | **description here**        |  |
-| rst-fact-retrieval-11b      | description here       |   |
-| rst-summarization-11b      | description here       |   |
-| rst-temporal-reasoning-11b      | description here       |   |
-| rst-information-extraction-11b      | description here       |   |
-| rst-intent-detection-11b      | description here       |   |
-| rst-topic-classification-11b      | description here       |   |
-| rst-word-sense-disambiguation-11b      | description here       |   |
-| rst-natural-language-inference-11b      | description here       |   |
-| rst-sentiment-classification-11b      | description here       |   |
-| rst-gaokao-rc-11b      | description here       |   |
-| rst-gaokao-cloze-11b      | description here       |   |
-| rst-gaokao-writing-11b      | description here       |   |
+| **rst-all-11b**                     | **Trained with all the signals below except signals that are used to train Gaokao models**        | **All applications below** |
+| rst-fact-retrieval-11b              | Trained with the following signals: WordNet meaning, WordNet part-of-speech, WordNet synonym, WordNet antonym, wikiHow category hierarchy, Wikidata relation, Wikidata entity typing, Paperswithcode entity typing       | Fact retrieval  |
+| rst-summarization-11b               | Trained with the following signals: DailyMail summary, Paperswithcode summary, arXiv summary, wikiHow summary    | Summarization  |
+| rst-temporal-reasoning-11b          | Trained with the following signals: DailyMail temporal information, wikiHow procedure       |  Temporal reasoning |
+| rst-information-extraction-11b      | Trained with the following signals: Paperswithcode entity, Paperswithcode entity typing, Wikidata entity typing, Wikidata relation, Wikipedia entity       |  Named entity recognition, relation extraction|
+| rst-intent-detection-11b            | Trained with the following signals: wikiHow goal-step relation       | Intent prediction  |
+| rst-topic-classification-11b        | Trained with the following signals: DailyMail category, arXiv category, wikiHow text category, Wikipedia section title | Topic classification   |
+| rst-word-sense-disambiguation-11b   | Trained with the following signals: WordNet meaning, WordNet part-of-speech, WordNet synonym, WordNet antonym   | Word sense disambiguation, part-of-speech tagging  |
+| rst-natural-language-inference-11b  | Trained with the following signals: ConTRoL dataset, DREAM dataset, LogiQA dataset, RACE & RACE-C dataset, ReClor dataset, DailyMail temporal information       | Natural language inference, multiple-choice question answering  |
+| rst-sentiment-classification-11b    | Trained with the following signals: Rotten Tomatoes sentiment, Wikipedia sentiment | Sentiment Classification  |
+| rst-gaokao-rc-11b                   | Trained with multiple-choice QA datasets that are used to train the [T0pp](https://huggingface.co/bigscience/T0pp) model       | Multiple-choice question answering, Gaokao reading comprehension  |
+| rst-gaokao-cloze-11b                | Trained with manually crafted cloze datasets    |  Cloze filling, Gaokao cloze questions |
+| rst-gaokao-writing-11b              | Trained with example essays from past Gaokao-English exams and grammar error correction signals       | Essay writing, grammar error correction  |