INDEX

Explanations

a specific coding style or formatting in programming-related content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

–

-0.34

 already

-0.33

애

-0.33

℠

-0.33

qJ

-0.32

–

-0.31

 otherwise

-0.31

AndroidJUnit

-0.30

“

-0.29

POSITIVE LOGITS

 Paglinawan

0.65

 esternos

0.63

 AssemblyCulture

0.63

iffance

0.62

 plufieurs

0.57

WriteTagHelper

0.54

genodigd

0.54

conomie

0.52

licone

0.52

 BytesLike

0.52

Activations Density 0.016%

a specific coding style or formatting in programming-related content

No Comments

No Known Activations