INDEX

Explanations

words related to administrative terms

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

*(

-0.29

 mentales

-0.29

 Mantel

-0.29

 substr

-0.28

########.

-0.28

 polí

-0.28

بيض

-0.28

StreamReader

-0.27

 verder

-0.27

Aktualisiert

-0.27

POSITIVE LOGITS

VersionUID

0.75

 BoxFit

0.66

хьтан

0.64

ніципа

0.64

AutoScaleMode

0.61

/**

0.60

InputBorder

0.57

localctx

0.57

UserScript

0.57

/*

0.54

Activations Density 0.001%

words related to administrative terms

No Comments

No Known Activations