INDEX

Explanations

instances of the word "through" or related phrases indicating a process or progression

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 10-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_10/width_16k/average_l0_113

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 kysy

-0.60

 doméstico

-0.52

 fós

-0.50

 zijne

-0.45

 sagrado

-0.42

 Segurança

-0.42

 henkil

-0.41

 sièges

-0.41

 cible

-0.40

ßerdem

-0.40

POSITIVE LOGITS

 through

1.96

Through

1.92

 Through

1.92

through

1.88

 THROUGH

1.72

THROUGH

1.52

 thru

1.40

 attraverso

1.37

 Thru

1.36

Thru

1.35

Activations Density 0.224%

instances of the word "through" or related phrases indicating a process or progression

No Comments

No Known Activations