INDEX

Explanations

terms related to modification and management of digital content and data

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

Actions that can be easily undone or reversed

np_acts-logits-general · gemini-2.0-flash

can be easily removed or revoked

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 سعی

-0.66

 hauptsächlich

-0.59

primarily

-0.59

embal

-0.58

 vooral

-0.56

 primarily

-0.56

 Mainly

-0.56

 głównie

-0.56

ibol

-0.54

 mainly

-0.54

POSITIVE LOGITS

 easily

1.62

 fácilmente

1.57

 facilmente

1.39

easily

1.39

 anytime

1.38

 facilement

1.29

 without

1.25

ได้

1.24

 kapan

1.10

 Easily

1.10

Activations Density 0.799%

terms related to modification and management of digital content and data

Actions that can be easily undone or reversed

can be easily removed or revoked

No Comments

No Known Activations