INDEX

Explanations

concepts related to identification and classification processes

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

Koordinaten

-0.40

 Sorge

-0.40

いかがでしたか

-0.37

ktes

-0.35

FORME

-0.35

籍

-0.35

 rhestr

-0.35

 Caution

-0.34

以上就是

-0.34

 ligiloj

-0.34

POSITIVE LOGITS

WEBPACK

0.68

UnusedPrivate

0.62

saraba

0.55

Personendaten

0.53

AISSEE

0.52

 للمعارف

0.50

nloa

0.49

enumii

0.47

]")]

0.47

currentColor

0.46

Activations Density 0.023%

concepts related to identification and classification processes

No Comments

No Known Activations