INDEX

Explanations

document structure indicators such as headers or sections marked by the "<bos>" token

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 16-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_16/width_65k/average_l0_128

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.16.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

 serio

-0.56

 EconPapers

-0.54

iastes

-0.52

diğinde

-0.50

انگلیسی

-0.50

 feedback

-0.49

CppCodeGen

-0.49

PageIndex

-0.49

 everybody

-0.49

 Waray

-0.49

POSITIVE LOGITS

principalTable

0.65

DockStyle

0.60

 négociations

0.58

WithIOException

0.57

makeConstraints

0.57

twimg

0.55

TestFixture

0.54

windowFixed

0.54

 cœurs

0.54

 therefrom

0.53

Activations Density 0.022%

document structure indicators such as headers or sections marked by the "<bos>" token

No Comments

No Known Activations