INDEX

Explanations

data-related terms and structures in a document

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

Chinese text and strings

np_acts-logits-general · gemini-2.0-flash

watchlist and collection management

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

PreferredItem

-0.81

 flavor

-0.81

 theater

-0.77

 theaters

-0.75

 flavored

-0.74

 Ее

-0.74

 Flavor

-0.73

ArrowToggle

-0.72

 parlor

-0.70

 behavior

-0.70

POSITIVE LOGITS

 apologise

0.99

 utilise

0.97

colour

0.97

colourful

0.96

Organis

0.95

Colour

0.94

 bahawa

0.93

COLOUR

0.93

 utilising

0.92

centre

0.91

Activations Density 0.075%

data-related terms and structures in a document

Chinese text and strings

watchlist and collection management

No Comments

No Known Activations