INDEX

Explanations

empathy for sadness or difficulties

When a turn ends (tokenizing signals like <end_of_turn> and assistant reply boundaries), i.e., it detects conversation turn boundaries—especially assistant/model response segments.

New Auto-Interp

Configuration

Prompts (Dashboard)

238,145 prompts, 512 tokens each

Dataset (Dashboard)

lmsys + oasst1

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 entusiasmo

0.66

สนุก

0.65

楽し

0.64

 enjoyment

0.63

 entusi

0.60

즐

0.60

楽しい

0.58

 enthusi

0.57

 enthous

0.57

 excitement

0.56

POSITIVE LOGITS

:(

0.55

:(

0.55

 Hope

0.52

Wishing

0.49

Sending

0.48

Hope

0.48

 Hang

0.48

😔

0.47

大丈夫

0.47

 cared

0.47

Activations Density 0.006%