INDEX

Explanations

references to writing and corresponding verbs in the context of letters or documents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 35-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_35/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.35.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

 být

-0.50

époque

-0.44

 mít

-0.42

Seen

-0.42

 spoken

-0.41

 beaten

-0.40

 być

-0.40

 congrès

-0.40

ğimiz

-0.40

摇头

-0.40

POSITIVE LOGITS

 took

1.32

 gave

1.20

did

1.20

 went

1.09

 drove

0.99

 drew

0.98

 began

0.95

 came

0.94

took

0.94

did

0.92

Activations Density 0.036%

references to writing and corresponding verbs in the context of letters or documents

No Comments

No Known Activations