INDEX

Explanations

the word "might" indicating speculation or possibility

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

 corporate

-0.75

Corporate

-0.68

 Corporate

-0.64

 CORPOR

-0.63

corporate

-0.62

Jeografia

-0.56

nodoc

-0.50

 corpor

-0.50

‌پدی

-0.49

 StatefulWidget

-0.47

POSITIVE LOGITS

 might

1.51

Might

1.47

 Might

1.41

might

1.39

 MIGHT

1.15

 Mighty

0.68

 podría

0.67

 mighty

0.65

mighty

0.64

 pourraient

0.64

Activations Density 0.028%

the word "might" indicating speculation or possibility

No Comments

No Known Activations