INDEX

Explanations

occurrences of characters or names, often with a focus on initials and titles

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

First initial of a name

np_acts-logits-general · gemini-2.0-flash

initials preceding names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

✨:

-0.64

testng

-0.54

qrstuvwxyz

-0.54

AndEndTag

-0.52

openqa

-0.50

writeString

-0.49

enaar

-0.47

ersdorf

-0.46

idjan

-0.45

abello

-0.45

POSITIVE LOGITS

AntiForgeryToken

0.59

SourceChecksum

0.57

 Skor

0.54

amik

0.54

__*/

0.52

 invokingState

0.50

 iParam

0.49

 Thane

0.48

Szk

0.47

 StatefulWidget

0.47

Activations Density 0.209%

occurrences of characters or names, often with a focus on initials and titles

First initial of a name

initials preceding names

No Comments

No Known Activations