INDEX

Explanations

elements related to dining or culinary experiences

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

places and nice experiences

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

 Wiktionnaire

-0.59

 kaarangay

-0.58

 propOrder

-0.57

bootstrapcdn

-0.56

 Meksiku

-0.52

enciaga

-0.46

 getService

-0.46

 gyhoeddwyd

-0.45

 dziecka

-0.44

 deſt

-0.44

POSITIVE LOGITS

 paradise

0.46

paradise

0.41

'{@

0.39

 Paradise

0.38

able

0.37

 sunny

0.37

 entertainment

0.36

cous

0.36

 sleeps

0.36

 Möglichkeiten

0.36

Activations Density 0.010%

elements related to dining or culinary experiences

places and nice experiences

No Comments

No Known Activations