INDEX

Explanations

terms related to subscription and receiving updates or news

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

subscribe or subscription

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

 juiz

-0.56

-0.48

PERMISSION

-0.47

Allen

-0.46

tough

-0.45

Hons

-0.45

QMetaType

-0.45

Rig

-0.45

damaged

-0.44

ferro

-0.44

POSITIVE LOGITS

 subscription

1.12

 subscriber

1.11

 subscribe

1.02

Subscriber

1.01

 subscribing

0.99

 subscribed

0.98

 Subscription

0.96

 Subscribe

0.95

 subscribers

0.94

 Subscriber

0.93

Activations Density 0.012%

terms related to subscription and receiving updates or news

subscribe or subscription

No Comments

No Known Activations