INDEX

Explanations

references to online articles and reports

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

citing research and texts

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

Personensuche

-0.60

adaptiveStyles

-0.56

AndEndTag

-0.52

 nahilalakip

-0.48

oa̍t

-0.45

 podjela

-0.45

Ɓ

-0.44

 CONSEQU

-0.44

 caoutchouc

-0.43

giu

-0.43

POSITIVE LOGITS

PAM

0.40

PAM

0.36

 OMITBAD

0.36

 Vinci

0.36

 Wals

0.36

tvguidetime

0.36

 cáo

0.35

bon

0.35

 Surgeons

0.34

Vinci

0.34

Activations Density 0.032%

references to online articles and reports

citing research and texts

No Comments

No Known Activations