INDEX

Explanations

technical and scientific terminologies related to neuroscience and research methodology

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_33/width_131k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.33.hook_resid_post

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

TagHelper

-0.50

ModelForm

-0.46

migrationBuilder

-0.45

ReusableCell

-0.41

__":

-0.38

gewiesen

-0.37

ValueStyle

-0.36

Portale

-0.36

Tracce

-0.36

 <<<<<<<<<<<<<<

-0.35

POSITIVE LOGITS

 ensures

1.47

 means

1.37

 makes

1.33

 allows

1.28

 enables

1.26

 membuatnya

1.23

使得

1.23

 precludes

1.22

 prevents

1.21

 sprawia

1.20

Activations Density 0.349%

technical and scientific terminologies related to neuroscience and research methodology

No Comments

No Known Activations