INDEX

Explanations

programming and data-related terms

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

ulement

-0.33

winston

-0.30

 zufolge

-0.28

 columnIndex

-0.28

 uLocal

-0.26

imp

-0.26

 foreground

-0.26

❉

-0.26

庁

-0.26

 translateY

-0.25

POSITIVE LOGITS

pushFollow

0.71

Personendaten

0.67

 EconPapers

0.65

IBarButtonItem

0.62

rungsseite

0.62

 nonUne

0.61

 ModelExpression

0.60

:✨

0.58

uxxxx

0.57

 Mikey

0.56

Activations Density 0.000%

programming and data-related terms

No Comments

No Known Activations