INDEX

Explanations

expressions of positive experiences and preferences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 queſta

-0.50

parian

-0.47

 Wicidata

-0.46

ecake

-0.46

aimeJ

-0.45

 Timberlake

-0.44

 Waſſer

-0.44

yons

-0.44

eking

-0.44

஢

-0.44

POSITIVE LOGITS

Kjelder

0.37

 addictive

0.36

 addicted

0.36

XtraEditors

0.35

 addic

0.34

 جد

0.33

 permanent

0.33

 stole

0.33

 kasarigan

0.32

adaptiveStyles

0.32

Activations Density 0.203%

expressions of positive experiences and preferences

No Comments

No Known Activations