INDEX

Explanations

quantities or numbers related to counts and variety

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

numbers or counts

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

ittoria

-0.46

Источники

-0.42

lamy

-0.41

 Schmitz

-0.41

<>

-0.40

 encima

-0.40

 ['',

-0.40

 setAddress

-0.40

 themselves

-0.39

 seeming

-0.39

POSITIVE LOGITS

GEBURTSDATUM

0.74

PerformLayout

0.60

 kasarigan

0.59

 ब्रेकडाउन

0.54

 typelib

0.54

 iprot

0.54

 &___

0.52

 Paglinawan

0.48

contentLoaded

0.47

 utafitiHapana

0.46

Activations Density 0.001%

quantities or numbers related to counts and variety

numbers or counts

No Comments

No Known Activations