INDEX

Explanations

sections related to summaries, profiles, and organizational approaches or guidelines

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

overview, ad, profile, zone, court

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

UnsafeEnabled

-0.41

lwjgl

-0.40

alamualaikum

-0.38

DoubleQuotes

-0.38

/**

-0.35

 RISERV

-0.34

Boxscore

-0.34

featureID

-0.34

oa̍t

-0.33

 “...

-0.33

POSITIVE LOGITS

colt

0.47

Word

0.47

 podjela

0.47

μβρίου

0.45

SequentialGroup

0.45

loyalty

0.45

Loyalty

0.43

Grace

0.43

handy

0.42

 Catalina

0.42

Activations Density 0.000%

sections related to summaries, profiles, and organizational approaches or guidelines

overview, ad, profile, zone, court

No Comments

No Known Activations