INDEX

Explanations

references to "watching" or related activities, particularly in a media or streaming context

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

"watching" or "observers"

np_acts-logits-general · gemini-2.0-flash

watching and observers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

Abp

-0.67

AddTagHelper

-0.66

teries

-0.60

◆◆

-0.59

ègre

-0.57

ArgsConstructor

-0.55

 potr

-0.54

 EdgeInsets

-0.54

 Ruskin

-0.53

 JAXBElement

-0.53

POSITIVE LOGITS

tower

0.94

 atenta

0.92

dogs

0.91

towers

0.83

 watches

0.79

 Watches

0.78

 WATCH

0.77

 closely

0.77

dog

0.76

 watchers

0.75

Activations Density 0.108%

references to "watching" or related activities, particularly in a media or streaming context

"watching" or "observers"

watching and observers

No Comments

No Known Activations