INDEX

Explanations

instances of the word "feature" and its variations

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

"feature" or "features"

np_acts-logits-general · gemini-2.0-flash

feature, features, featured, featuring

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 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

 Dryden

-0.53

 Anthropological

-0.53

 Anmeldung

-0.52

Sko

-0.52

querque

-0.52

Answered

-0.51

 Gonçalves

-0.50

Discography

-0.49

 Malone

-0.49

 Northampton

-0.49

POSITIVE LOGITS

 Feature

1.27

 feature

1.25

feature

1.22

Feature

1.09

 FEATURE

1.05

 features

1.03

features

0.99

FEATURE

0.98

 Features

0.95

 featured

0.93

Activations Density 0.038%

instances of the word "feature" and its variations

"feature" or "features"

feature, features, featured, featuring

No Comments

No Known Activations