INDEX

Explanations

the word "here" in various contexts

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

 here

-0.71

principalTable

-0.68

execSQL

-0.64

)");

-0.64

 IndexPath

-0.64

YNAM

-0.63

Personensuche

-0.62

GOTREF

-0.62

ViewById

-0.61

Expedia

-0.61

POSITIVE LOGITS

bootstrapcdn

0.63

 fausse

0.57

 său

0.56

 finnas

0.51

kript

0.50

 chi̍t

0.49

SHARE

0.48

 محفوظة

0.47

ameters

0.47

Selanjutnya

0.46

Activations Density 0.042%

the word "here" in various contexts

No Comments

No Known Activations