INDEX

Explanations

phrases indicating the presentation of lists or resources

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 27-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_27/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.27.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tagHelperRunner

-0.77

featureID

-0.75

aarrggbb

-0.74

 مشين

-0.72

 arşivlendi

-0.68

 disambiguazione

-0.67

хьтан

-0.62

HasColumnName

-0.62

postsleuth

-0.62

 pinulongan

-0.61

POSITIVE LOGITS

 five

1.02

 seven

0.98

 eight

0.93

six

0.88

 nine

0.88

ten

0.84

 four

0.82

 three

0.79

 zehn

0.78

 diez

0.75

Activations Density 0.125%

phrases indicating the presentation of lists or resources

No Comments

No Known Activations