INDEX

Explanations

instances of the word "in" to indicate location or context

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

 zoude

-1.11

 dezelve

-0.87

 desmotivaciones

-0.78

 geda

-0.74

arangay

-0.72

 zijne

-0.71

 pungkasnya

-0.70

 verantwoorde

-0.69

 RIPRODUZIONE

-0.68

 plufieurs

-0.67

POSITIVE LOGITS

in

1.06

in

0.99

In

0.92

In

0.90

 into

0.76

getIn

0.74

IN

0.71

IN

0.71

 isIn

0.63

isIn

0.62

Activations Density 0.005%

instances of the word "in" to indicate location or context

No Comments

No Known Activations