INDEX

Explanations

instances of code block beginnings or specific formatting indicators in programming languages

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_4/width_65k/average_l0_66

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.4.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

subject

-0.53

MLLoader

-0.49

rather

-0.48

Literatuur

-0.47

 szín

-0.45

 zwłaszcza

-0.44

 subject

-0.44

sowie

-0.42

least

-0.42

 rather

-0.41

POSITIVE LOGITS

 հղումներ

0.90

__*/

0.88

 lenker

0.88

Personendaten

0.80

 StatelessWidget

0.77

ThroughAttribute

0.77

WaitGroup

0.76

abestanden

0.75

setopt

0.75

 estekak

0.73

Activations Density 0.053%

instances of code block beginnings or specific formatting indicators in programming languages

No Comments

No Known Activations