INDEX

Explanations

references to individuals associated with specific roles or positions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 giornal

-0.49

허

-0.48

 Secretaries

-0.45

 Carriers

-0.45

altezza

-0.44

 xung

-0.43

Encyklopedia

-0.43

 fédéral

-0.43

Przypisy

-0.42

 päivä

-0.42

POSITIVE LOGITS

 former

0.94

 Englishman

0.86

 youngster

0.86

 veteran

0.84

 Irishman

0.83

Former

0.83

 Frenchman

0.82

 midfielder

0.81

former

0.79

 Former

0.75

Activations Density 0.165%

references to individuals associated with specific roles or positions

No Comments

No Known Activations