INDEX

Explanations

pieces of code related to processing CPU IDs and match length calculations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 17-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_17/width_65k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.17.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

SOUNDBITE

-0.62

WithIOException

-0.57

XDECREF

-0.54

+#+

-0.54

 زاده

-0.52

tvguidetime

-0.51

 статистики

-0.49

壤

-0.48

 éve

-0.48

 configureStore

-0.47

POSITIVE LOGITS

 kasarigan

0.66

 propOrder

0.65

enumii

0.64

 esternos

0.64

SourceChecksum

0.62

Diwedd

0.57

 Италијани

0.57

Tembelea

0.57

كويكب

0.56

TagMode

0.56

Activations Density 0.032%

pieces of code related to processing CPU IDs and match length calculations

No Comments

No Known Activations