INDEX

Explanations

phrases related to numerical sequences or calculations

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

ValueStyle

-0.76

CppMethod

-0.69

Datuak

-0.64

oredCriteria

-0.60

 lenker

-0.59

 kasarigan

-0.57

ſſung

-0.56

存于互联网档案馆

-0.56

yntaxException

-0.55

 iNdEx

-0.54

POSITIVE LOGITS

<eos>

0.38

Concluding

0.38

 münd

0.36

 LIRE

0.35

Greetings

0.34

 féminine

0.34

kuu

0.33

Herzliche

0.32

活

0.32

Contract

0.31

Activations Density 0.013%

phrases related to numerical sequences or calculations

No Comments

No Known Activations