INDEX

Explanations

words and phrases associated with diverse contexts and actions, particularly in entertainment, health, and relationships

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

specific terms and entities

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

HtmlAttribute

-0.34

 noqa

-0.33

MessageState

-0.32

queles

-0.30

 שוליים

-0.30

 bVar

-0.30

DPA

-0.30

minecraftforge

-0.30

 empe

-0.30

θει

-0.30

POSITIVE LOGITS

uxxxx

0.63

hyrchwyd

0.60

ViewImports

0.60

bitField

0.56

 resourceCulture

0.56

Personensuche

0.55

 Infórmanos

0.55

setVerticalGroup

0.52

NameInMap

0.52



0.52

Activations Density 0.137%

words and phrases associated with diverse contexts and actions, particularly in entertainment, health, and relationships

specific terms and entities

No Comments

No Known Activations