INDEX

Explanations

formatted text elements and their properties

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

formatting and layout

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

ordeel

-0.47

yarnpkg

-0.44

npmjs

-0.43

 själva

-0.43

deville

-0.41

usehen

-0.41

cookieParser

-0.41

ungsbedingungen

-0.40

uete

-0.40

odor

-0.39

POSITIVE LOGITS

̈́

0.60

 propOrder

0.55

ArgsConstructor

0.42

Personendaten

0.41

AxisAlignment

0.41

 Taktlose

0.41

 ExecuteAsync

0.41

0.40

 мәкал

0.39

}-*/;

0.39

Activations Density 0.105%

formatted text elements and their properties

formatting and layout

No Comments

No Known Activations