INDEX

Explanations

adjectives describing intensity or quality

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

degree or quality with an attribute

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

 adentro

-0.50

 varandra

-0.49

ammans

-0.49

 juos

-0.47

뀜

-0.46

 afuera

-0.45

 föruts

-0.44

 träd

-0.44

AllAfrica

-0.44

 écri

-0.43

POSITIVE LOGITS

хьтан

0.62

(&:

0.50

 autorytatywna

0.50

 noDo

0.46

ModelForm

0.45

 Wicidata

0.45

ViewInit

0.44

OfClass

0.42

OF

0.42

IVEREF

0.40

Activations Density 0.045%

adjectives describing intensity or quality

degree or quality with an attribute

No Comments

No Known Activations