INDEX

Explanations

references to personal experiences or identities

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

 Inscrivez

-0.75

 wireType

-0.74

 kyse

-0.57

 huelga

-0.56

 maquilla

-0.55

 acestea

-0.55

 convite

-0.54

 Diffraction

-0.54

 Gottesdienst

-0.53

rachtet

-0.52

POSITIVE LOGITS

 personal

0.96

personal

0.96

 Personal

0.91

Personal

0.90

 PERSONAL

0.78

ersonal

0.75

PERSONAL

0.74

individual

0.71

 individual

0.70

Individual

0.68

Activations Density 0.013%

references to personal experiences or identities

No Comments

No Known Activations