INDEX

Explanations

instances of inappropriate or abusive behavior

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

PO

-0.34

 evidence

-0.32

SAV

-0.30

de

-0.29

-0.28

mer

-0.28

 turns

-0.28

POSITIVE LOGITS

 ویکی‌پدی

0.77

Autoritní

0.73

 ModelExpression

0.73

setVerticalGroup

0.72

 protoimpl

0.72

 NSCoder

0.71

 ddelweddau

0.69

adaptiveStyles

0.68

WebElementEntity

0.68

ParallelGroup

0.67

Activations Density 0.065%

instances of inappropriate or abusive behavior

No Comments

No Known Activations