INDEX

Explanations

references to specific sections or results related to findings in a study or paper

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_0/width_65k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.0.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

 Boven

-0.57

webpack

-0.57

kelt

-0.57

 Shil

-0.56

ط

-0.55

 Lohn

-0.53

 Rosh

-0.53

nele

-0.53

 Reco

-0.53

Kec

-0.52

POSITIVE LOGITS

kozó

0.75

VI

0.68

 suivantes

0.66

 sahiptir

0.66

 <=",

0.65

 desee

0.63

 للاسماء

0.63

 suivante

0.62

///<

0.62

igslist

0.60

Activations Density 0.017%

references to specific sections or results related to findings in a study or paper

No Comments

No Known Activations