INDEX

Explanations

references to specific terms or categories that are definitively defined or recognized

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.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

出版年

-0.48

tagez

-0.44

 ekos

-0.44

 circonst

-0.43

chieht

-0.43

ulto

-0.42

 adulthood

-0.40

ydable

-0.40

 nakalista

-0.40

Дереккөздер

-0.39

POSITIVE LOGITS

 certain

1.65

 Certain

1.61

certain

1.58

Certain

1.57

 CERTAIN

1.30

 sure

1.11

 certaine

1.11

sure

0.99

 certo

0.98

 ciertos

0.96

Activations Density 0.092%

references to specific terms or categories that are definitively defined or recognized

No Comments

No Known Activations