INDEX

Explanations

expressions indicating judicial or governmental criticism and accountability

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

 незавершена

-0.57

PropertyGroup

-0.43

WireFormatLite

-0.41

belea

-0.39

 Krü

-0.37

 Kimura

-0.36

 fine

-0.36

wegs

-0.35

 CreateTagHelper

-0.35

Vac

-0.34

POSITIVE LOGITS

 itſelf

0.59

rungsseite

0.59

 laſt

0.53

 Majefty

0.53

 esternos

0.52

 fubject

0.52

TagMode

0.51

enderror

0.51

 himſelf

0.51

 Jefus

0.50

Activations Density 0.009%

expressions indicating judicial or governmental criticism and accountability

No Comments

No Known Activations