INDEX

Explanations

terms associated with spiritual concepts like salvation and redemption

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

FontOfSize

-0.49

깐

-0.47

 layouts

-0.47

 Fancy

-0.46

+#+#

-0.45

Kik

-0.44

 beards

-0.44

 bull

-0.44

fif

-0.44

joba

-0.44

POSITIVE LOGITS

 salvation

0.83

 salvación

0.81

 Salvation

0.75

WebElementEntity

0.66

 sauver

0.59

TagMode

0.57

 keselamatan

0.57

 køb

0.56

 térmico

0.54

 nahilalakip

0.52

Activations Density 0.003%

terms associated with spiritual concepts like salvation and redemption

No Known Activations