INDEX

Explanations

expressions that indicate opinion or assertion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

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

 chrétien

-0.50

 bénéficiaire

-0.43

 étranger

-0.40

 enví

-0.39

 gobern

-0.38

 complejos

-0.37

 couvercle

-0.36

 ennemi

-0.36

 désert

-0.36

 informacion

-0.36

POSITIVE LOGITS

WriteTagHelper

0.70

hamcrest

0.66

 NavController

0.66

ConstraintMaker

0.63

ffindor

0.60

 okuyayım

0.60

 Taktlose

0.59

ſſung

0.59

ElementException

0.59

terdam

0.59

Activations Density 0.012%

expressions that indicate opinion or assertion

No Comments

No Known Activations