INDEX

Explanations

terms associated with legal arguments and procedural elements

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

 يتيمه

-0.82

 تضيفلها

-0.68

PyExc

-0.63

 Chwiliwch

-0.63

SBATCH

-0.62

出版年

-0.62

OuterAlt

-0.60

XmlEnum

-0.57

XmlAccessType

-0.57

 esternos

-0.57

POSITIVE LOGITS

 própria

0.33

 Menschheit

0.31

 changing

0.31

 wide

0.31

setViewportView

0.31

cept

0.31

gania

0.30

 uncertain

0.30

ukone

0.30

own

0.29

Activations Density 0.061%

terms associated with legal arguments and procedural elements

No Comments

No Known Activations