INDEX

Explanations

references to television shows and broadcasts

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

television

np_acts-logits-general · gemini-2.0-flash

TV followed by career, icon, ad, or series

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 lenker

-0.66

RemoveField

-0.61

]")]

-0.52

AlterField

-0.52

ارنة

-0.52

twimg

-0.51

 متعلقه

-0.51

 ProtoMessage

-0.50

cognito

-0.49

AndEndTag

-0.49

POSITIVE LOGITS

 stations

0.70

 estekak

0.67

 programmes

0.61

ButtonModule

0.60

 Stations

0.60

 cameras

0.60

0.59

脚注の使い方

0.56

 towers

0.56

 station

0.55

Activations Density 0.053%

references to television shows and broadcasts

television

TV followed by career, icon, ad, or series

No Comments

No Known Activations