test with new data

README.md

@@ -12,18 +12,19 @@ tags:
 - NLG
 - evaluation
 size_categories:
-- 100K<n<1M
+- 1M<n<10M
 language:
 - en
 ---
-# 🚢  Stanford Human Preferences Dataset (SHP)
+# 🚢  Stanford Human Preferences Dataset v2 (SHP-2)
 
 ## Summary
 
-SHP is a dataset of **385K collective human preferences** over responses to questions/instructions in 18 different subject areas, from cooking to legal advice.
+SHP-2 is a dataset of **4.8M collective human preferences** over responses to questions/instructions in 129 different subject areas, from cooking to legal advice. It is an extended version of the original 385K [SHP dataset](https://huggingface.co/datasets/stanfordnlp/SHP).
+
 The preferences are meant to reflect the helpfulness of one response over another, and are intended to be used for training RLHF reward models and NLG evaluation models (e.g., [SteamSHP](https://huggingface.co/stanfordnlp/SteamSHP-flan-t5-xl)).
 
-Each example is a Reddit post with a question/instruction and a pair of top-level comments for that post, where one comment is more preferred by Reddit users (collectively). 
+Each example is a Reddit or StackExchange post with a question/instruction and a pair of top-level comments for that post, where one comment is more preferred by Reddit / StackExchange users (collectively). 
 SHP exploits the fact that if comment A was written *after* comment B but has a higher score nonetheless, then A is ostensibly more preferred to B.
 If A had been written before B, then we could not conclude this, since its higher score could have been the result of more visibility.
 We chose data where the preference label is intended to reflect which response is more *helpful* rather than which is less *harmful*, the latter being the focus of much past work.
@@ -33,7 +34,7 @@ Most notably, all the data in SHP is naturally occurring and human-written, wher
 
 | Dataset              | Size | Input                      | Label                       | Domains           | Data Format                           | Length |
 | -------------------- | ---- | -------------------------- | ---------------------------- | ------------------------- | ------------------------------------- | --------------- |
-| SHP                  | 385K | Naturally occurring human-written responses   | Collective Human Preference  | 18 (labelled)  | Question/Instruction + Response (Single-turn) | up to 10.1K T5 tokens |
+| SHP-2                  | 4.8M | Naturally occurring human-written responses   | Collective Human Preference  | 129 (labelled)  | Question/Instruction + Response (Single-turn) | up to 10.1K T5 tokens |
 | HH-RLHF    | 91K | Dialogue with LLM  | Individual Human Preference  | not labelled   | Live Chat (Multi-turn)              | up to 1.5K T5 tokens |
 
 How is SHP different from other datasets that have scraped Reddit, like [ELI5](https://huggingface.co/datasets/eli5#source-data)?
@@ -42,7 +43,7 @@ It also contains data from more domains:
 
 | Dataset              | Size | Comments + Scores  | Preferences  | Number of Domains  |
 | -------------------- | ---- | ------------------ | -------------| ------------------ |
-| SHP                  | 385K | Yes                | Yes          | 18  |
+| SHP-2                  | 4.8M | Yes                | Yes          | 129 (70 from Reddit, 59 from StackExchange)  |
 | ELI5                 | 270K | Yes                | No           | 3   |
 
 
@@ -55,13 +56,13 @@ Here's how to get the data using Huggingface's `datasets` library:
 from datasets import load_dataset
 
 # Load all the data 
-dataset = load_dataset("stanfordnlp/shp")
+dataset = load_dataset("stanfordnlp/shp-2")
 
 # Load one of the subreddits
-dataset = load_dataset("stanfordnlp/shp", data_dir="askculinary")
+dataset = load_dataset("stanfordnlp/shp-2", data_dir="reddit/askculinary")
 ```
 
-Here's an example from `askculinary/train.json`:
+Here's an example from `reddit/askculinary/train.json`:
 ```
 {
     `post_id`:"qt3nxl",

stackexchange/stack_academia/test.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77b7d82c625463433f11cb14ddcdc817b85e328d20fecbf4d7933e79973db49b
+size 1067151

stackexchange/stack_academia/train.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3694433d50610a1c653537f2c104dcbbcaaf51958afd8a525c206e6bc3ea1e6
+size 25012875

stackexchange/stack_academia/validation.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:73055bdbe492a5c55a5d02c20369301b19822f98658da5c0a7fbd944e4c6c935
+size 1232052

stackexchange/stack_android/test.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3c9dbdaa75d0d13026dc28c18913d179a751dc39ecd546331a09386a4882b42
+size 195287

stackexchange/stack_android/train.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d3e3b9bcfa9181f4603f4c049541b29fbbfb9e0bf0b6b7a6577f1031202397a
+size 2180355

stackexchange/stack_android/validation.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc099f0d2867b6093333544705e04bdf14a4cc7d546630b8067d5bba57adf0d6
+size 109010

stackexchange/stack_apple/test.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:823f5da9ba57a5d9221f2daaf862ce26aefd8f1c7bf752a7a2bafb255c559c40
+size 824978

stackexchange/stack_apple/train.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36363f786b1b9f52837b6fb6a866b3ef299f56f4dea5b4ffaa2a9a1a8d163378
+size 14326265

stackexchange/stack_apple/validation.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b6ba9fdfaca6d5a3cc0d461e4ce0ef6bcd9ce976e6aed4d3247105a77833ba6
+size 718357

stackexchange/stack_arduino/test.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0519fa305efb42ee21bf731bb6a5d9f22eadabbf32bc9e417bbca82fc2b5c614
+size 59120

stackexchange/stack_arduino/train.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:069f41674d9161ae4aee4bba024f505142824fd3c1188faa50e4837d0d3c738f
+size 1229291

stackexchange/stack_arduino/validation.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f99cc582bebd61c8f0d82720b31efaa6ba9493afb2ff5d50cb502b41defb488
+size 73433

stackexchange/stack_askubuntu/test.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14521550427f3a20c57fd8cc214cb8d697e62bb83f6fb1b5d8a22f4c5aede53a
+size 1703214

stackexchange/stack_askubuntu/train.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f58c543417082042b46cd9bdb93a8c8bbf5645a46e12a6c76dea6963018f91b4
+size 30860762

stackexchange/stack_askubuntu/validation.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1744d6760cc9eae8d9450564fb8993700d033f7e51937836f99b2e60e927d87
+size 1718593

stackexchange/stack_aviation/test.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e052a7aa41dd70aac7af66ab91a4e1e25f60586f8dcbd131d6be0aaae40e23b
+size 285006

stackexchange/stack_aviation/train.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c70e25d740547aff739fc9feb8df5269e87d2cecfcd48086601ec2178da7689c
+size 7031848

stackexchange/stack_aviation/validation.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36ead4c5492700ddeb5bbec8cf75088b1f173c492a3ac26160a5b9daccdad62a
+size 360754