INDEX

Explanations

terms related to data, specifically focusing on the concept of information

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_32/width_131k/average_l0_88

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.32.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

fjspx

-0.59

ValueStyle

-0.48

endpush

-0.46

 Inscrivez

-0.46

 bleue

-0.45

 gesprek

-0.44

脚注の使い方

-0.43

endphp

-0.42

zeß

-0.42

ScopeManager

-0.41

POSITIVE LOGITS

FormBorderStyle

0.44

 Parthen

0.42

SCI

0.42

 paran

0.40

 infra

0.40

 BIBSYS

0.40

drawSprites

0.40

GAG

0.39

cch

0.39

 vito

0.39

Activations Density 0.001%

terms related to data, specifically focusing on the concept of information

No Comments

No Known Activations