INDEX

Explanations

proper nouns and names

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

names followed by comma

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

lovakia

-0.49

ulemon

-0.47

 Justiz

-0.43

 palpable

-0.43

isissez

-0.43

granate

-0.43

anjutan

-0.42

jillo

-0.42

 crescere

-0.42

 Brasília

-0.42

POSITIVE LOGITS

KommentareTeilen

0.53

 Funeral

0.52

WillAppear

0.49

Autoritní

0.49

ModelSerializer

0.47

 informée

0.46

Abitanti

0.45

 funeral

0.45

 funer

0.44

StoryboardSegue

0.42

Activations Density 0.001%

proper nouns and names

names followed by comma

No Comments

No Known Activations