INDEX

Explanations

numeric or URL references related to meetings and recordings

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

meeting recording identifiers

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

AntiForgeryToken

-0.46

appé

-0.44

 Wiktionnaire

-0.42

 HttpNotFound

-0.40

THISDAY

-0.39

 OMITTED

-0.38

YSTEM

-0.38

上角

-0.38

 Infórmanos

-0.36

atypes

-0.35

POSITIVE LOGITS

StoryboardSegue

0.54

 дописавши

0.44

Kaynakça

0.44

featureID

0.41

\#:

0.40

хьтан

0.40

bildēt

0.40

 Veranstal

0.40

 accessible

0.39

VersionUID

0.39

Activations Density 0.003%

numeric or URL references related to meetings and recordings

meeting recording identifiers

No Comments

No Known Activations