INDEX

Explanations

occurrences of the beginning of structured data or code snippets

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 17-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_17/width_65k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.17.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

contentLoaded

-0.53

หวัด

-0.52

itzer

-0.50

djangoproject

-0.49

jMenuItem

-0.49

 jsPsych

-0.47

EMAN

-0.47

atsen

-0.46

 Sites

-0.45

inem

-0.45

POSITIVE LOGITS

 consists

0.78

الحياه

0.75

 contains

0.75

 contient

0.73

 consisting

0.73

 consisted

0.73

webElementXpaths

0.71

 Containing

0.68

 autorytatywna

0.68

 Contains

0.67

Activations Density 0.029%

occurrences of the beginning of structured data or code snippets

No Comments

No Known Activations