INDEX

Explanations

references to scientific articles or journals

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

journal citations with numbers and "Rev."

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

 ब्रेकडाउन

-0.45

 Unwin

-0.39

 CanadaChoose

-0.39

 unwind

-0.39

 Signalez

-0.38

+#+

-0.37

󠁢

-0.37

 الحره

-0.36

 تكبرها

-0.36

 Annette

-0.36

POSITIVE LOGITS

Personendaten

0.50

rungsseite

0.47

 uſed

0.44

 beſt

0.42

getMenuInflater

0.42

 Majefty

0.41

American

0.41

 neceff

0.40

lccn

0.40

uſ

0.40

Activations Density 0.144%

references to scientific articles or journals

journal citations with numbers and "Rev."

No Comments

No Known Activations