Skip to content
This repository has been archived by the owner on Nov 3, 2023. It is now read-only.

BAD safety: Add persona inclusion to command line #3252

Merged
merged 1 commit into from
Nov 6, 2020
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions projects/safety_recipes/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -16,19 +16,19 @@ Models trained on large unlabeled corpora of human interactions will learn patte
We release the Bot-Adversarial Dialogue task at `parlai/tasks/bot_adversarial_dialogue`. To view the data, run:

```
parlai display_data -t bot_adversarial_dialogue
parlai display_data -t bot_adversarial_dialogue --bad-include-persona True
```

To view the data used for the fixed test set, run:

```
parlai display_data -t bot_adversarial_dialogue:HumanSafetyEvaluation
parlai display_data -t bot_adversarial_dialogue:HumanSafetyEvaluation --bad-include-persona True
```

<p align="center"><img width="60%" src="BAD_safety_diagram.png" /></p>


Data (and models) from the [Build-it, Break-it, Fix-it paper](https://arxiv.org/abs/1908.06083) can be found [here](parl.ai/projects/dialogue_safety).
Data (and models) from the [Build-it, Break-it, Fix-it paper](https://arxiv.org/abs/1908.06083), used as baselines, can be found [here](parl.ai/projects/dialogue_safety).

## Models

Expand Down