INDEX

Explanations

references to addresses and numerical data

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

numbers and punctuation

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.24.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 selfies

-0.60

 autorytatywna

-0.59

 '\\;'

-0.58

 incentiv

-0.57

 للمعارف

-0.53

 selfie

-0.53

 utafitiHapana

-0.53

expandindo

-0.52

StructEnd

-0.52

Cyfeiriadau

-0.51

POSITIVE LOGITS

faßt

0.62

 Einfluß

0.52

 aDecoder

0.49

 Bewußt

0.48

 muß

0.43

 dentaire

0.42

 daß

0.41

 Saddam

0.41

vskip

0.41

BrowserModule

0.39

Activations Density 0.010%

references to addresses and numerical data

numbers and punctuation

No Comments

No Known Activations