INDEX

Explanations

programming-related keywords and syntax

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_18/width_65k/average_l0_88

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.18.hook_mlp_out

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

 Coin

-0.52

Verdana

-0.50

tomic

-0.48

 minced

-0.48

extend

-0.46

 Gegenteil

-0.45

 Peque

-0.45

ället

-0.44

impact

-0.44

 Klas

-0.44

POSITIVE LOGITS

twimg

0.77

Hochspringen

0.71

 AssemblyTitle

0.67

 فريبيس

0.65

///</

0.64

StoryboardSegue

0.64

ConstraintMaker

0.62

Pautan

0.60

featureID

0.59

存于互联网档案馆

0.59

Activations Density 0.025%

programming-related keywords and syntax

No Comments

No Known Activations