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This repository has been archived by the owner on Nov 3, 2023.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

Changed blenderbot stats to the right ones and created feedback section #3865

Merged
merged 1 commit into from
Jul 26, 2021

Conversation

liliarose
Copy link
Contributor

@liliarose liliarose commented Jul 26, 2021

  • Accidentally used blenderbot90M instead of blenderbot_3B (blenderbot 1) for comparison. (checked blenderbot2 & that's correct; should be since it was autogenerated instead of hand-added)
  • Added a feedback section

Notes:

/private/home/wzhang4343/ParlAI/tmp/blender90M (old/wrong one)

{"Response to Offensive Language:Affirmation Detection": {
    "Sentiment analysis (% positive sentiment)": 0.4196078431372549, 
    "Negation detection (% without negatives)": 0.013725490196078431, 
    "Safety classifier (% flagged offensive)": 0.22745098039215686}}

/private/home/wzhang4343/ParlAI/blenderbot_3B/response_to_offensive_language_metrics.json

{"Response to Offensive Language:Affirmation Detection": 
    {"Sentiment analysis (% positive sentiment)": 0.6294117647058823, 
        "Negation detection (% without negatives)": 0.06470588235294118, 
        "Safety classifier (% flagged offensive)": 0.2529411764705882}}

/private/home/wzhang4343/ParlAI/parlai_internal/sample_model_cards/kshuster/safety_bench_res/response_to_offensive_language_metrics.json (blenderbot2)

{"Response to Offensive Language:Affirmation Detection": 
    {"Sentiment analysis (% positive sentiment)": 0.9882352941176471, 
        "Negation detection (% without negatives)": 0.9823529411764705, 
        "Safety classifier (% flagged offensive)": 0.045098039215686274}}

@liliarose liliarose merged commit 173bb9f into master Jul 26, 2021
@liliarose liliarose deleted the liliarose-patch-2 branch July 26, 2021 22:56
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants