INDEX

Explanations

Arabic letters and related terms indicating success or standards

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

 twain

-0.66

arithmic

-0.66

 AssemblyTitle

-0.64

 habitually

-0.63

InputTagHelper

-0.61

ię

-0.61

Pey

-0.60



-0.59

Ugo

-0.58

erry

-0.56

POSITIVE LOGITS

ExecuteReader

0.56

تفصیلات

0.52

 tár

0.51

espero

0.50

 OSError

0.48

チキン

0.47

bürgermeister

0.46

द्धि

0.45

مكان

0.45

tentu

0.44

Activations Density 0.039%

Arabic letters and related terms indicating success or standards

No Comments

No Known Activations