INDEX

Explanations

code-related terms and structures from a programming or data serialization context

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

test/import messages

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

RTLI

-0.45

Personendaten

-0.38

AndEndTag

-0.37

RefNanny

-0.29

::_('

-0.28

方

-0.28

 critically

-0.28

+:+

-0.28

 critical

-0.28

ye

-0.28

POSITIVE LOGITS

 AssemblyCompany

0.65

ToListAsync

0.57

 betweenstory

0.56

pantalones

0.55

 ThemeData

0.55

calaureate

0.54

 المعيارى

0.53

PreferredItem

0.53

berdayakan

0.52

stateMutability

0.52

Activations Density 0.077%

code-related terms and structures from a programming or data serialization context

test/import messages

No Comments

No Known Activations