INDEX

Explanations

specific names or terms related to people, particularly surnames

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 1-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_1/width_16k/average_l0_69

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.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

:✨

-0.68

endphp

-0.67

 myſelf

-0.57

 himſelf

-0.56

 itſelf

-0.52

httphttps

-0.48

/*

-0.46

 ſein

-0.44

featureID

-0.44

 avoit

-0.44

POSITIVE LOGITS

ander

0.76

ANDER

0.66

 aspiration

0.63

 mundane

0.59

 aspirations

0.58

 boring

0.55

 bored

0.54

ει

0.53

 ambition

0.51

 aspire

0.51

Activations Density 0.400%

specific names or terms related to people, particularly surnames

No Comments

No Known Activations