INDEX

Explanations

phrases that convey reassurance and mitigate fear

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

words like "worry," "fear," or negations

np_acts-logits-general · gemini-2.0-flash

don't worry

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

 sportback

-0.60

 maji

-0.54

 niyang

-0.53

aragus

-0.53

 Stonehenge

-0.51

 unspeakable

-0.50

########.

-0.50

 slavery

-0.49

 airspace

-0.49

 móg

-0.49

POSITIVE LOGITS

 worry

1.64

 fear

1.43

 Worry

1.42

 worries

1.38

worry

1.28

Fear

1.27

 Fear

1.24

don

1.23

 fret

1.23

fear

1.22

Activations Density 0.176%

phrases that convey reassurance and mitigate fear

words like "worry," "fear," or negations

don't worry

No Comments

No Known Activations