INDEX

Explanations

technical terms and functions used in programming or data manipulation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 1-gemmascope-mlp-16k

Configuration

google/gemma-scope-2b-pt-mlp/layer_1/width_16k/average_l0_105

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.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

 TimeUnit

-0.73

 BorderSide

-0.69

 FetchType

-0.69

 RouterModule

-0.68

 SDLK

-0.68

endforeach

-0.66

 MainAxisSize

-0.66

WriteLiteral

-0.66

 useHistory

-0.65

JNIEnv

-0.65

POSITIVE LOGITS

 igång

0.57

 distanciation

0.52

DockStyle

0.49

 varandra

0.45

 älskar

0.43

 stří

0.43

 bör

0.43

าก็

0.43

As

0.43

</caption>

0.43

Activations Density 0.605%

technical terms and functions used in programming or data manipulation

No Comments

No Known Activations