INDEX

Explanations

phrases that indicate relationships, especially those involving specificity or definitions

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

which describes or states

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

aarrggbb

-0.68

Спасылкі

-0.54

GOTREF

-0.54

 برانيه

-0.53

Geplaatst

-0.52

providedIn

-0.52

 mergeFrom

-0.52

BeginContext

-0.50

indd

-0.50

InjectAttribute

-0.50

POSITIVE LOGITS

 Wochenende

0.43

 fermée

0.42

 nocturno

0.42

 tecnológica

0.41

 mußte

0.41

picos

0.39

bases

0.39

 Económica

0.39

 žen

0.39

 FRANCISCO

0.38

Activations Density 0.244%

phrases that indicate relationships, especially those involving specificity or definitions

which describes or states

No Comments

No Known Activations