INDEX

Explanations

phrases that indicate reasoning or justifications

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_32k/average_l0_76

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

float32

Hook Name

blocks.12.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

 المعيارى

-0.94

JsonHelper

-0.78

ParallelGroup

-0.66

featureID

-0.65

IUrlHelper

-0.62

帖最后由

-0.60

SOUNDBITE

-0.59

্ষ

-0.59

getDescriptor

-0.58

Datuak

-0.58

POSITIVE LOGITS

OGND

0.66

atrième

0.53

XtraReports

0.50

BagLayout

0.49

MPC

0.46

MPC

0.45

QUAD

0.45

LETON

0.45

 buffed

0.44

 brewers

0.43

Activations Density 0.126%

phrases that indicate reasoning or justifications

No Comments

No Known Activations