INDEX

Explanations

phrases indicating ownership or coverage by particular entities

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

 Komunikasi

-0.42

RegressionTest

-0.42

CloseOperation

-0.42

 сре

-0.41

OGND

-0.41

 AssemblyCompany

-0.40

 MenuView

-0.40

 ever

-0.40

 ſou

-0.39

évaluateur

-0.39

POSITIVE LOGITS

withIdentifier

0.52

 referenties

0.47

 resourceCulture

0.45

 lenker

0.43

elemField

0.43

 armadura

0.40

 ligiloj

0.40

 kandungan

0.40

>::

0.40

Abraço

0.39

Activations Density 0.101%

phrases indicating ownership or coverage by particular entities

No Comments

No Known Activations