Showing a random sample of {N_COMMENTS} comments (out of {set_length} comments)
{:else}
Showing all {set_length} comments
{/if}
{#if show_step2_info}
Interpreting this visualization
This plot has the same layout as the All Topics visualization, but now, each box in this plot represents an individual comment that belongs to your selected topic area.
The x-axis represents our prediction of your toxicity rating for each comment (we'll call these "your ratings")
The left side (white background) is the Non-toxic side (comments that'll be allowed to remain)
The right side (grey background) is the Toxic side (comments that will be deleted)
Comment boxes are plotted along the x-axis based on our prediction of your toxicity rating for that comment
The color of the box indicates the system's rating for the same comment; you may want to focus on the red-colored boxes that indicate disagreements between "your ratings" and the system's ratings
{/if}
{/if}
{#if show_checkboxes}
Comments
{/if}
{#if show_checkboxes}
Numerical ratings:
ShowHide
{#if use_model}
Our prediction of your decision + ratings:
ShowHide
{/if}
{/if}
{#key evidence}
Comment
{#if show_your_decision_ratings}
Our prediction of your decision
{#if show_num_ratings}
Our prediction of your rating
{/if}
{/if}
System decision
{#if show_checkboxes}
White = Non-toxic, Grey = Toxic
{/if}
{#if show_num_ratings}
System rating
{/if}
{#if show_checkboxes}
{#if use_model}
Potential error type
{#if show_checkboxes}
Darker red = Greater potential system error
{/if}
Potential toxicity categories
{/if}
{/if}
Do you agree with the system?
{#if !show_checkboxes}
Remove
{/if}
{#if show_checkboxes}
Add Evidence
{/if}
{#each slice as item (item.id + table_id)}
{item.comment}
{#if show_your_decision_ratings}
{item.user_decision}
{#if show_num_ratings}
{item.user_rating}
{/if}
{/if}
{item.system_decision}
{#if show_num_ratings}
{item.system_rating}
{/if}
{#if show_checkboxes}
{#if use_model}
{item.error_type}
{item.toxicity_category}
{/if}
{/if}
of your decision
of your rating
decision
{#if show_checkboxes} White = Non-toxic,
Grey = Toxic {/if}
rating
type
{#if show_checkboxes} Darker red = Greater
potential system error {/if}
categories
with the system?
Evidence