INDEX

Explanations

phrases related to recommendations for viewing or using content

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

Followed by verbs related to utility or observation

np_acts-logits-general · gemini-2.0-flash

watch, use, seeing, follow

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

Autoritní

-0.60

to

-0.58

bootstrapcdn

-0.55

umburg

-0.50

 sufic

-0.49

ensement

-0.49

 suficientes

-0.49

httphttps

-0.49

 Enough

-0.48

existsSync

-0.45

POSITIVE LOGITS

 للمعارف

0.76

ReusableCell

0.69

 مشين

0.66

tvguidetime

0.65

URIComponent

0.61

 OnInit

0.60

 sessionFactory

0.60

 ולה

0.59

 oneself

0.59

IntoConstraints

0.58

Activations Density 0.324%

phrases related to recommendations for viewing or using content

Followed by verbs related to utility or observation

watch, use, seeing, follow

No Comments

No Known Activations