INDEX

Explanations

words and phrases that suggest quantifiable outcomes or comparisons

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

snapshot of, lot worse, fact that

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

ⓧ

-0.66

 snippetHide

-0.47

Autoritní

-0.47

 Chwiliwch

-0.46

KommentareTeilen

-0.45

 CanadaChoose

-0.45

RegistryLite

-0.43

-0.42

 habet

-0.41

AnchorTagHelper

-0.41

POSITIVE LOGITS

bootstrapcdn

0.51

 GenerationType

0.50

essentiel

0.44

 ModelExpression

0.43

 Vordergrund

0.42

ieteur

0.42

 thưởng

0.41

Kaieteur

0.38

 Schluß

0.37

 Balkon

0.37

Activations Density 0.080%

words and phrases that suggest quantifiable outcomes or comparisons

snapshot of, lot worse, fact that

No Comments

No Known Activations