INDEX

Explanations

words associated with brightness and illumination

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Jacob

-0.55

Raptor

-0.52

Etape

-0.51

Jacob

-0.50

 Maverick

-0.48

 Margo

-0.47

⑵

-0.46

ViewGroup

-0.44

 Pace

-0.44

 Raptor

-0.44

POSITIVE LOGITS

Shine

0.96

 Shine

0.93

 shine

0.91

shine

0.91

hine

0.76

 shines

0.74

shining

0.73

 shining

0.68

 SHIN

0.66

 Sunshine

0.66

Activations Density 0.003%

words associated with brightness and illumination

No Comments

No Known Activations