INDEX

Explanations

instances of the word "work."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 3-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_3/width_16k/average_l0_90

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.3.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

 >=",

-0.54

 Svensson

-0.44

TagMode

-0.43

sillon

-0.43

ʚ

-0.42

ellido

-0.42

ghis

-0.40

endregion

-0.40

__::

-0.39

cipais

-0.38

POSITIVE LOGITS

Work

3.53

 Work

3.52

work

3.36

 WORK

3.03

WORK

3.00

 work

2.89

 Works

2.14

ワーク

2.06

Works

2.06

工作

2.06

Activations Density 0.153%

instances of the word "work."

No Comments

No Known Activations