INDEX

Explanations

instances of written content or references to writing and authorship

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

references to writing tasks or requirements

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

writing or creating documents

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-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

 écoute

-0.67

ReusableCell

-0.62

録音

-0.55

 OMIT

-0.54

//-->

-0.52

 linkovi

-0.51

AnimationsModule

-0.51

 Matériau

-0.51

ฟัง

-0.51

iNdEx

-0.51

POSITIVE LOGITS

 script

0.74

 letter

0.68

 scripts

0.65

 screenplay

0.65

 about

0.64

 history

0.62

script

0.61

 letters

0.59

 blog

0.59

 finis

0.57

Activations Density 0.051%

instances of written content or references to writing and authorship

references to writing tasks or requirements

writing or creating documents

No Comments

No Known Activations