INDEX

Explanations

affirmative responses and agreement phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

 tartalomajánló

-0.76

Efq

-0.63

 blessés

-0.60

 comprennent

-0.57

tinyos

-0.56

 cortesía

-0.55

 againſt

-0.55

 uſed

-0.55

 Theſe

-0.53

ölkerung

-0.52

POSITIVE LOGITS

Yes

1.17

yes

1.14

Yes

1.10

yes

1.07

Oui

0.91

YES

0.91

YES

0.90

Yep

0.86

Yep

0.84

بله

0.80

Activations Density 0.150%

affirmative responses and agreement phrases

No Comments

No Known Activations

affirmative responses and agreement phrases

No Comments

No Known Activations