INDEX

Explanations

geographical locations or significant places

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

DockStyle

-0.63

makeConstraints

-0.47

 Cæsar

-0.45

AccessorTable

-0.44

 inverse

-0.44

 nuke

-0.43

full

-0.43

 propOrder

-0.42

 dimensionless

-0.42

Grüsse

-0.42

POSITIVE LOGITS

RegressionTest

0.48

netbeans

0.46

 médicale

0.42

 paille

0.41

 hiér

0.41

 mxArray

0.40

cerpt

0.40

 ویکی‌پدی

0.39

+#+

0.37

новниш

0.36

Activations Density 0.028%

geographical locations or significant places

No Comments

No Known Activations