Report for lxyuan/distilbert-base-multilingual-cased-sentiments-student
#89
by
giskard-bot
- opened
Hi Team,
This is a report from Giskard Bot Scan 🐢.
We have identified 8 potential vulnerabilities in your model based on an automated scan.
This automated analysis evaluated the model on the dataset tweet_eval (subset sentiment, split validation).
👉Performance issues (1)
| Vulnerability | Level | Data slice | Metric | Transformation | Deviation |
|---|---|---|---|---|---|
| Performance | medium 🟡 | text contains "friday" |
Precision = 0.432 | — | -7.05% than global |
🔍✨Examples
For records in the dataset where `text` contains "friday", the Precision is 7.05% lower than the global Precision.| text | label | Predicted label |
|
|---|---|---|---|
| 27 | every time I hear alright by Kendrick I think it's j Cole's Black Friday | neutral | positive (p = 0.49) |
| 38 | ##$$## Black Friday Deals Olympus OM-D E-M5 Digital Camera - Black - with Olympus 12-50mm f/3.5-5.6 EZ Zoom Lens - B... | neutral | positive (p = 0.58) |
| 144 | When niggas in the bus are playing Kendrick and Cole's Black Friday out loud >>>>>> | neutral | negative (p = 0.45) |
👉Ethical issues (1)
| Vulnerability | Level | Data slice | Metric | Transformation | Deviation |
|---|---|---|---|---|---|
| Ethical | medium 🟡 | — | Fail rate = 0.071 | Switch Religion | 6/85 tested samples (7.06%) changed prediction after perturbation |
🔍✨Examples
When feature “text” is perturbed with the transformation “Switch Religion”, the model changes its prediction in 7.06% of the cases. We expected the predictions not to be affected by this transformation.| text | Switch Religion(text) | Original prediction | Prediction after perturbation | |
|---|---|---|---|---|
| 171 | I really want to thank Angela Merkel for letting all those refugees in. She shows what humanity is about. May God bless her #refugeeswelcome | I really want to thank Angela Merkel for letting all those refugees in. She shows what humanity is about. May allah bless her #refugeeswelcome | positive (p = 0.60) | negative (p = 0.51) |
| 808 | @user may you be blessed by guns, god and hungry wet holes before Scott Walker builds his border wall and Donald Trump sends you home!" | @user may you be blessed by guns, allah and hungry wet holes before Scott Walker builds his border wall and Donald Trump sends you home!" | positive (p = 0.73) | negative (p = 0.55) |
| 1256 | Sitting on my luggage and smelling Jerusalem \u002cclick like if you hope to visit Israel soon . December tour... | Sitting on my luggage and smelling kumbh mela \u002cclick like if you hope to visit Israel soon . December tour... | negative (p = 0.48) | positive (p = 0.41) |
👉Robustness issues (5)
| Vulnerability | Level | Data slice | Metric | Transformation | Deviation |
|---|---|---|---|---|---|
| Robustness | major 🔴 | — | Fail rate = 0.393 | Transform to uppercase | 393/1000 tested samples (39.3%) changed prediction after perturbation |
🔍✨Examples
When feature “text” is perturbed with the transformation “Transform to uppercase”, the model changes its prediction in 39.3% of the cases. We expected the predictions not to be affected by this transformation.| text | Transform to uppercase(text) | Original prediction | Prediction after perturbation | |
|---|---|---|---|---|
| 1074 | I'm so frustrated with Game of Thrones and I'm only on the 10th episode | I'M SO FRUSTRATED WITH GAME OF THRONES AND I'M ONLY ON THE 10TH EPISODE | negative (p = 0.88) | positive (p = 0.54) |
| 1816 | Guys... I'm seriously... #Stonehill right now... unranked and beating #3 #NewHaven in the 4th quarter... CBS College Sports... | GUYS... I'M SERIOUSLY... #STONEHILL RIGHT NOW... UNRANKED AND BEATING #3 #NEWHAVEN IN THE 4TH QUARTER... CBS COLLEGE SPORTS... | negative (p = 0.70) | positive (p = 0.43) |
| 1681 | """Why America May Go To Hell""- wish it wouldve been completed and i wish i could read the contents of it... by MLK" | """WHY AMERICA MAY GO TO HELL""- WISH IT WOULDVE BEEN COMPLETED AND I WISH I COULD READ THE CONTENTS OF IT... BY MLK" | negative (p = 0.42) | positive (p = 0.52) |
| Vulnerability | Level | Data slice | Metric | Transformation | Deviation |
|---|---|---|---|---|---|
| Robustness | major 🔴 | — | Fail rate = 0.307 | Transform to title case | 307/1000 tested samples (30.7%) changed prediction after perturbation |
🔍✨Examples
When feature “text” is perturbed with the transformation “Transform to title case”, the model changes its prediction in 30.7% of the cases. We expected the predictions not to be affected by this transformation.| text | Transform to title case(text) | Original prediction | Prediction after perturbation | |
|---|---|---|---|---|
| 1816 | Guys... I'm seriously... #Stonehill right now... unranked and beating #3 #NewHaven in the 4th quarter... CBS College Sports... | Guys... I'M Seriously... #Stonehill Right Now... Unranked And Beating #3 #Newhaven In The 4Th Quarter... Cbs College Sports... | negative (p = 0.70) | positive (p = 0.43) |
| 1681 | """Why America May Go To Hell""- wish it wouldve been completed and i wish i could read the contents of it... by MLK" | """Why America May Go To Hell""- Wish It Wouldve Been Completed And I Wish I Could Read The Contents Of It... By Mlk" | negative (p = 0.42) | positive (p = 0.65) |
| 99 | omg then I sat on my floor in front of the TV and bawled over Shawn when he was performing on that one show | Omg Then I Sat On My Floor In Front Of The Tv And Bawled Over Shawn When He Was Performing On That One Show | negative (p = 0.51) | positive (p = 0.48) |
| Vulnerability | Level | Data slice | Metric | Transformation | Deviation |
|---|---|---|---|---|---|
| Robustness | major 🔴 | — | Fail rate = 0.144 | Transform to lowercase | 144/1000 tested samples (14.4%) changed prediction after perturbation |
🔍✨Examples
When feature “text” is perturbed with the transformation “Transform to lowercase”, the model changes its prediction in 14.4% of the cases. We expected the predictions not to be affected by this transformation.| text | Transform to lowercase(text) | Original prediction | Prediction after perturbation | |
|---|---|---|---|---|
| 1812 | Chuck Norris cut off his left nut and donated it to science. You may know it as Jupiter. | chuck norris cut off his left nut and donated it to science. you may know it as jupiter. | positive (p = 0.47) | negative (p = 0.44) |
| 226 | Good morning...back after a couple of days off for Labor Day weekend. So today is my Monday. I make no promises. Ready with @user | good morning...back after a couple of days off for labor day weekend. so today is my monday. i make no promises. ready with @user | neutral (p = 0.42) | positive (p = 0.52) |
| 1499 | UK: Chancellor Osborne try to sneak into 1st class train with standard ticket - Breaking News Buzz | uk: chancellor osborne try to sneak into 1st class train with standard ticket - breaking news buzz | positive (p = 0.43) | negative (p = 0.42) |
| Vulnerability | Level | Data slice | Metric | Transformation | Deviation |
|---|---|---|---|---|---|
| Robustness | major 🔴 | — | Fail rate = 0.137 | Add typos | 137/1000 tested samples (13.7%) changed prediction after perturbation |
🔍✨Examples
When feature “text” is perturbed with the transformation “Add typos”, the model changes its prediction in 13.7% of the cases. We expected the predictions not to be affected by this transformation.| text | Add typos(text) | Original prediction | Prediction after perturbation | |
|---|---|---|---|---|
| 1886 | "Randy Orton was the first 3rd generation superstar, Natalya was the first 3rd generation diva. When are we going for the 4th generation?" | "Randy Orton was the first 3rd generation spuersftwr, Natzalya was the first 3rd generation duav. When are we going for the 4th gemefation?" | positive (p = 0.60) | negative (p = 0.48) |
| 1229 | @user I installed Madden 16 Deluxe last Monday night for PS4 and still haven't received my packs today nor the reward for opening 50 | @user I ijstalled Madden 16 Deluxe last Mondxy night for LX4 and still haven't receives mt packs today mnor the reward for opening %50 | neutral (p = 0.43) | negative (p = 0.36) |
| 1251 | "2,780,589 people could have seen 'Rahul Gandhi in Bengaluru' since its 1st mention until it became a Trending Topic. #trndnl" | "2,780,589 peopoe could hafe seen 'Raul Gandhi un Bengaluru' since its 1st mdntion until it became a Trending GTopic. #trndnl" | positive (p = 0.55) | negative (p = 0.47) |
| Vulnerability | Level | Data slice | Metric | Transformation | Deviation |
|---|---|---|---|---|---|
| Robustness | medium 🟡 | — | Fail rate = 0.092 | Punctuation Removal | 92/1000 tested samples (9.2%) changed prediction after perturbation |
🔍✨Examples
When feature “text” is perturbed with the transformation “Punctuation Removal”, the model changes its prediction in 9.2% of the cases. We expected the predictions not to be affected by this transformation.| text | Punctuation Removal(text) | Original prediction | Prediction after perturbation | |
|---|---|---|---|---|
| 1489 | Curtis Painter...we have a chance again! Can't believe Kerry Collins didn't throw us a pick-six tonight | Curtis Painter we have a chance again Can t believe Kerry Collins didn t throw us a pick six tonight | negative (p = 0.41) | neutral (p = 0.42) |
| 1971 | Bowling tomorrow c; Don\u2019t want things to be awkard lol | Bowling tomorrow c Don\u2019t want things to be awkard lol | positive (p = 0.40) | negative (p = 0.40) |
| 1952 | @user @user Yellow journalism. But you know? This may be Harper's Waterloo | @user @user Yellow journalism But you know This may be Harper s Waterloo | negative (p = 0.42) | positive (p = 0.42) |
👉Overconfidence issues (1)
| Vulnerability | Level | Data slice | Metric | Transformation | Deviation |
|---|---|---|---|---|---|
| Overconfidence | medium 🟡 | avg_digits(text) < 0.011 |
Overconfidence rate = 0.291 | — | +18.82% than global |
🔍✨Examples
For records in the dataset where `avg_digits(text)` < 0.011, we found a significantly higher number of overconfident wrong predictions (183 samples, corresponding to 29.09% of the wrong predictions in the data slice).| text | avg_digits(text) | label | Predicted label |
|
|---|---|---|---|---|
| 1900 | Monsanto wants to merge with Syngenta and change name to wash away the bad reputation (3rd most disliked company!): | 0.00869565 | neutral | negative (p = 0.95) |
| neutral (p = 0.03) | ||||
| 1503 | @user this is absolutely ridiculous. Wasn't it just ""national ice cream day"" on Sunday? Who makes up these days? This isn't official. This is" | 0 | neutral | negative (p = 0.95) |
| neutral (p = 0.04) | ||||
| 1203 | I may have just mentally assembled the most insane conspiracy web about the Dr. Luke / Kesha sitch. | 0 | neutral | negative (p = 0.94) |
| neutral (p = 0.04) |
Checkout out the Giskard Space and test your model.
Disclaimer: it's important to note that automated scans may produce false positives or miss certain vulnerabilities. We encourage you to review the findings and assess the impact accordingly.