INDEX

Explanations

references to the "Mad" series or characters related to it

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

owulf

-0.38

 getBy

-0.38

 Niro

-0.38

PLC

-0.37

trino

-0.37

glyphicon

-0.36

 Pili

-0.36

///

-0.35

 Punto

-0.34

Reina

-0.34

POSITIVE LOGITS

Mad

1.05

MAD

0.93

Mad

0.92

mad

0.88

 Maddox

0.88

MAD

0.86

 Madd

0.84

 Madsen

0.84

mad

0.79

 Maddie

0.73

Activations Density 0.004%

references to the "Mad" series or characters related to it

No Comments

No Known Activations