INDEX

Explanations

references to measurement or numerical values

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

names and places

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

 lenker

-0.87

 يتيمه

-0.87

Tembelea

-0.87

 pinulongan

-0.82

Datuak

-0.82

 Италијани

-0.76

 Wikimedijinoj

-0.75

 Normdatei

-0.74

 disambiguazione

-0.74

الحياه

-0.72

POSITIVE LOGITS

wußt

0.31

 fédé

0.28

égard

0.26

appré

0.25

setFrom

0.24

daß

0.24

Furthermore

0.23

require

0.23

 avoient

0.23

 tercih

0.23

Activations Density 0.064%

references to measurement or numerical values

names and places

No Comments

No Known Activations