INDEX

Explanations

references to numerical values or citations in scientific contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 17-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_17/width_65k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.17.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

vė

-0.46

गा

-0.44

-0.43

bey

-0.42

 detachable

-0.41

ėjo

-0.40

lick

-0.40

 Declaration

-0.40

trzy

-0.40

ën

-0.39

POSITIVE LOGITS

rungsseite

1.08

MigrationBuilder

0.91

Portale

0.90

 autorytatywna

0.84

 للمعارف

0.81

 تضيفلها

0.80

ValueStyle

0.79

 المعيارى

0.78

yntaxException

0.77

MessageTagHelper

0.76

Activations Density 0.077%

references to numerical values or citations in scientific contexts

No Comments

No Known Activations