sentence
stringlengths 4
1.48k
| label
stringclasses 3
values | source
stringclasses 3
values | split
stringclasses 1
value |
---|---|---|---|
Those 2 drinks are part of the HK culture and has years of history. It is so bad. | negative | dynasent_r2 | train |
I was told by the repair company that was doing the car repair that fixing the rim was "impossible" and to replace it. | negative | dynasent_r1 | train |
It is there to give them a good time . | neutral | sst_local | train |
Like leafing through an album of photos accompanied by the sketchiest of captions . | negative | sst_local | train |
Johnny was a talker and liked to have fun. | positive | dynasent_r1 | train |
It as burnt to a crisp black flavorless | negative | dynasent_r2 | train |
I called Moveaholics Jason was amazing he even offered to come out that night. | positive | dynasent_r1 | train |
Is this place expensive? | neutral | dynasent_r1 | train |
It's likely crowded at the busier times so keep that in mind. | neutral | dynasent_r1 | train |
I was just looking at bottom line price and my old Fiat 500 Pop Convertible to be paid off, since I was already pre-approved for a car loan with my credit union. | neutral | dynasent_r1 | train |
So many times I've passed by this cafe and not noticed it. | neutral | dynasent_r1 | train |
When I saw someone in a golf cart I asked him. | neutral | dynasent_r1 | train |
Seriously just fall off the bone. | positive | dynasent_r1 | train |
These guys don't care for the costumers at all. | negative | dynasent_r2 | train |
I loved this place. | positive | dynasent_r1 | train |
You can see all the helpers behind the scenes. | neutral | dynasent_r1 | train |
I swear! | neutral | dynasent_r1 | train |
The car was short. | negative | dynasent_r2 | train |
There are no little structures for dogs to go through, basically it's just a plot of grassy land for the dogs to run around in. | negative | dynasent_r1 | train |
Definitely use their green/garlic paste at the front. | positive | dynasent_r1 | train |
You could tell homemade. | positive | dynasent_r1 | train |
After all, Vegas has always had an element of kitsch in its public image. | neutral | dynasent_r1 | train |
I highly recommend any location but his. | positive | dynasent_r1 | train |
Then through out the party she didn't fill up pitchers for the kids and didn't listen to me when I asked her to put everyone's order in our bill. | negative | dynasent_r1 | train |
They call me at dinnertime. | neutral | dynasent_r1 | train |
I'll take H & H over these any day. | positive | dynasent_r1 | train |
this time around we skipped combo and ordered individual meat dishes
portions were similar but you probably do save few bucks doing combo. | positive | dynasent_r1 | train |
Nothing frozen. | neutral | dynasent_r1 | train |
This is a film tailor-made for those who when they were in high school would choose the Cliff-Notes over reading a full-length classic . | negative | sst_local | train |
The second time it was a bigger guy wth a shaved head who provided excellent customer service. | positive | dynasent_r1 | train |
Cons: it's true about the "Law & Order" or "CSI" looking walk to the strip from this hotel. | negative | dynasent_r1 | train |
he was the toughest player on the team | neutral | dynasent_r2 | train |
Feels like it's been around for too long | neutral | dynasent_r2 | train |
They are just as good at "soft skills" as translating. | positive | dynasent_r1 | train |
But it's not. | neutral | dynasent_r1 | train |
Which NEVER happens!! | neutral | dynasent_r1 | train |
But I also didn't go there to be impressed with fancy decor. | negative | dynasent_r1 | train |
We tried a new place. We had a seat at the bar and enjoyed two delicious breakfast cocktails. | positive | dynasent_r2 | train |
The wait time can be a little long (I've never waited less than 1h and never more than 2h30) and the maître D has been known to be a little rude when he's overwhelmed with clients. | negative | dynasent_r1 | train |
An exhausting family drama about a porcelain empire and just as hard a flick as its subject matter . | neutral | sst_local | train |
Wish they would sell the Stetson Chopped and ship it to me in San Diego... | negative | dynasent_r1 | train |
The dry-rub were well coated with rub and the buffalo were doused in sauce. | positive | dynasent_r1 | train |
I am starting to wonder about Wyndham. | neutral | dynasent_r1 | train |
The food is consistently delicious. | positive | dynasent_r2 | train |
Chowder chock full of clams. | positive | dynasent_r1 | train |
Nothing more than a run-of-the-mill action flick . | neutral | sst_local | train |
There are a lot of donut shops in Pittsburgh. | neutral | dynasent_r1 | train |
You could go to an oyster restaurant but you'll easily pay triple the price. | neutral | dynasent_r1 | train |
I have never in my life needed a scaling and root planning, it's always been an adult prophy which is a huge price difference (not only for me, but the office really makes out!). | neutral | dynasent_r1 | train |
I was thinking that the spaying was going to come out to be at least $100 (she was in heat), but by surprise there was an additional $15 for meds- which is totally understandable but it just caught me off guard since this is my first time getting a dog spayed- I just wished in addition of telling me that there is an additional charge when they're in heat that they would've told me that they would also give meds with another charge. | negative | dynasent_r1 | train |
It's surprising that this place has an A+ rating on the BBB. | negative | dynasent_r1 | train |
The theme was frozen! | neutral | dynasent_r1 | train |
The fact that the ` best part ' of the movie comes from a 60-second homage to one of Demme 's good films does n't bode well for the rest of it . | negative | sst_local | train |
Appreciate you being open when you say you will be open. | positive | dynasent_r1 | train |
And the great thing is that there are tons of other people who have the same feelings about this place. | positive | dynasent_r1 | train |
This is not a large space, and is the smaller of the two locations. | neutral | dynasent_r1 | train |
As well as I payed for the more higher one. | neutral | dynasent_r1 | train |
I told him that it was fine and to please do it anyway. | neutral | dynasent_r1 | train |
I looked for Linus but realized it is a little early... | neutral | dynasent_r1 | train |
I have left my dog with Melody for as long as 30 days. | positive | dynasent_r1 | train |
Full of bland hotels , highways , parking lots , with some glimpses of nature and family warmth , Time Out is a discreet moan of despair about entrapment in the maze of modern life . | negative | sst_local | train |
Very pleasant experience. | positive | dynasent_r2 | train |
They were not stingy about it like most Thai restaurants. | positive | dynasent_r1 | train |
And guess what, they show up fast in nice clean cars and are friendly. | positive | dynasent_r1 | train |
I could not bend over to lie on a standard patient table. | neutral | dynasent_r1 | train |
A marvel of production design . | positive | sst_local | train |
We could've used some French bread at end to soak up the remaining juice. | negative | dynasent_r1 | train |
Went this morning for breakfast . | neutral | dynasent_r1 | train |
They insist a review of the car Monday or Tuesday (being that it was late Friday).tiring | negative | dynasent_r2 | train |
We wanted to try this place cause it is not far from the office, and the menu looked decent. | positive | dynasent_r1 | train |
I had a look at the menu with my family. | neutral | dynasent_r1 | train |
The best thing about it was that at the end of my meal I went up to the register to pay the difference between the spaghetti and the pizza and they told me it was on them and apologized for the inconvenience. | positive | dynasent_r1 | train |
Not too many seats to fight over in this joint. | negative | dynasent_r2 | train |
It says anyway. | positive | dynasent_r1 | train |
This is now were Maharaja used to be. | neutral | dynasent_r1 | train |
#5: Spending the day at the pool, you can get nice overpriced cocktail drinks from Bouchon, but relieve your bankruptcy pain with the thought that you just live once and you better enjoy your life to the fullest, by ordering $12 iceplended margaritas and raspberry caipirinha in plastic cups. | negative | dynasent_r1 | train |
We came here for a big group lunch during a weekend wedding in AZ. | neutral | dynasent_r1 | train |
They say to stay for the desert but I do not agree. | negative | dynasent_r2 | train |
This was our first time trying a Korean style BBQ place. | neutral | dynasent_r1 | train |
One thing to note: this is primarily a take-out establishment. | neutral | dynasent_r1 | train |
The location is terrible. | negative | dynasent_r2 | train |
I'd come to the point in my life where I didn't want to beg and bribe friends to help me move. | neutral | dynasent_r1 | train |
I spent more than 6K on furniture at this particular Living Spaces, they deliver all damage furniture to me the first time. | negative | dynasent_r1 | train |
I ordered chips and guacamole. | neutral | dynasent_r1 | train |
Since we were flying home the next day...the TSA folks X-rayed the cake for hidden objects...I kept thinking..hands off our cake! | neutral | dynasent_r1 | train |
The place is small and getting a table is HARD, the wait for the food is a little longer than I'd like. | negative | dynasent_r1 | train |
If you have ok to good credit you shouldn't have issues. | positive | dynasent_r1 | train |
Yes, we enjoyed some appetizers and our bread, but we really just kept raving about the steaks. | positive | dynasent_r1 | train |
I know the two companies get mixed up in reviews so remember Excellent = Todd and Becca from Scottsdale Segway Tours. | positive | dynasent_r1 | train |
When I arrived there was only one person in line in front of me and I was very pleased as I was starving and in a hurry. | positive | dynasent_r1 | train |
Am I politically incorrect, like calling a flight attendant a stewardess?) | neutral | dynasent_r1 | train |
I found them at their Ft Mill location a few years back, but they were out of the way for me. | neutral | dynasent_r1 | train |
the food is so hot to handle and taste | negative | dynasent_r2 | train |
I bought a small weekend subscription for the 2017-2018 season which was about 7 shows. | neutral | dynasent_r1 | train |
For $18, for get half a fried chicken and sides! | positive | dynasent_r1 | train |
If you go a regular Korean al-a-carte restaurant, you will pay more. | neutral | dynasent_r1 | train |
Not to mention they leave their rolls outside the back door even when raining. | negative | dynasent_r1 | train |
Believe me- if you get stomach aches from dairy your going to be converted! | positive | dynasent_r1 | train |
The tamale was not too sweet, but enough to take the edge off the spice on the pork. | positive | dynasent_r1 | train |
It 's clear why Deuces Wild , which was shot two years ago , has been gathering dust on MGM 's shelf . | negative | sst_local | train |
Dataset Card for Sentiment Merged (SST-3, DynaSent R1, R2)
This is a dataset for 3-way sentiment classification of reviews (negative, neutral, positive). It is a merge of Stanford Sentiment Treebank (SST-3) and DynaSent Rounds 1 and 2.
Dataset Details
The SST-3, DynaSent R1, and DynaSent R2 datasets were randomly mixed to form a new dataset with 102,097 Train examples, 5,421 Validation examples, and 6,530 Test examples. See Table 1 for the distribution of labels within this merged dataset.
Table 1: Label Distribution for the Merged Dataset
Split | Negative | Neutral | Positive |
---|---|---|---|
Train | 21,910 | 49,148 | 31,039 |
Validation | 1,868 | 1,669 | 1,884 |
Test | 2,352 | 1,829 | 2,349 |
Table 2: Contribution of Sources to the Merged Dataset
Dataset | Samples | Percent (%) |
---|---|---|
DynaSent R1 Train | 80,488 | 78.83 |
DynaSent R2 Train | 13,065 | 12.80 |
SST-3 Train | 8,544 | 8.37 |
Total | 102,097 | 100.00 |
Dataset Description
SST-5 is the Stanford Sentiment Treebank 5-way classification (positive, somewhat positive, neutral, somewhat negative, negative). To create SST-3 (positive, neutral, negative), the 'somewhat positive' class was merged and treated as 'positive'. Similarly, the 'somewhat negative class' was merged and treated as 'negative'.
DynaSent is a sentiment analysis dataset and dynamic benchmark with three classification labels: positive, negative, and neutral. The dataset was created in two rounds. First, a RoBERTa model was fine-tuned on a variety of datasets including SST-3, IMBD, and Yelp. They then extracted challenging sentences that fooled the model, and validated them with humans. For Round 2, a new RoBERTa model was trained on similar (but different) data, including the Round 1 dataset. The Dynabench platform was then used to create sentences written by workers that fooled the model.
It’s worth noting that the source datasets all have class imbalances. SST-3 positive and negative are about twice the number of neutral. In DynaSent R1, the neutral are more than three times the negative. And in DynaSent R2, the positive are more than double the neutral. Although this imbalance may be by design for DynaSent (to focus on the more challenging neutral class), it still represents an imbalanced dataset. The risk is that the model will learn mostly the dominant class. Merging the data helps mitigate this imbalance. Although there is still a majority of neutral examples in the training dataset, the neutral to negative ratio in DynaSent R1 is 3.21, and this is improved to 2.24 in the merged dataset.
Another potential issue is that the models will learn the dominant dataset, which is DynaSent R1. See Table 2 for a breakdown of the sources contributing to the Merged dataset.
- Curated by: Jim Beno
- Language(s) (NLP): English
- License: MIT
Dataset Sources
- Repository: jbeno/sentiment
- Paper: Pending
Uses
The dataset is intended to be used for 3-way sentiment classification of reviews (negative, neutral, positive).
Dataset Structure
There are three CSV files: train_all.csv, val_all.csv, test_all.csv. Each represents the merged train, validation and test splits as defined by the original source datasets.
Column | Description |
---|---|
sentence | The review sentence |
label | The class label: negative, neutral, or positive |
source | The source dataset: sst_local, dyansent_r1, or dynasent_r2 |
split | The split: train, validation, or test |
Dataset Creation
Curation Rationale
The dataset was created to fine-tune models on sentiment classification. The idea was to create a diverse 3-way sentiment classification dataset with challenging reviews.
Source Data
See Stanford Sentiment Treebank and DynaSent for details.
Dataset Card Contact
- Downloads last month
- 32