INDEX

Explanations

phrases indicating belonging or membership within a group or community

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

being a part of

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

Jeografia

-0.48

 botines

-0.45

 dugaan

-0.43

 öffnungszeiten

-0.41

 następnie

-0.40

 jueces

-0.40

 wikipagina

-0.40

 miniaturka

-0.39

 Grüsse

-0.39

 lhes

-0.38

POSITIVE LOGITS

ValueStyle

0.67

join

0.66

 join

0.63

joining

0.60

JOIN

0.59

 JOIN

0.59

 joining

0.58

 joins

0.58

 joined

0.57

 Join

0.57

Activations Density 0.009%

phrases indicating belonging or membership within a group or community

being a part of

No Comments

No Known Activations