INDEX

Explanations

references to the sizes of data structures and memory allocation functions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

ungeon

-0.49

 they

-0.43

matite

-0.41

arium

-0.41

credential

-0.41

instrumental

-0.40

ereum

-0.39

út

-0.39

TextBoxColumn

-0.39

 They

-0.38

POSITIVE LOGITS

 sizeof

1.38

sizeof

1.35

setof

0.71

offsetof

0.69

nameof

0.65

typeof

0.63

 typeof

0.62

ISupport

0.61

 autorytatywna

0.61

 nameof

0.57

Activations Density 0.098%

references to the sizes of data structures and memory allocation functions

No Comments

No Known Activations