INDEX

Explanations

header tags indicating structured content in a document

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

 betweenstory

-0.96

 Италијани

-0.77

出版年

-0.75



-0.68

parsedMessage

-0.67

libft

-0.66

 Italijani

-0.66

TypedDataSet

-0.66

dafx

-0.66

sizeCache

-0.66

POSITIVE LOGITS

 Titel

0.40

 tunik

0.38

 Tief

0.38

«

0.36

setTitle

0.34

 http

0.34

 titel

0.34

 https

0.33

https

0.33

Title

0.33

Activations Density 0.001%

header tags indicating structured content in a document

No Comments

No Known Activations