INDEX

Explanations

parentheses in the text

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

while loops

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

 Biôgrafia

-0.56

󠁴

-0.55

 الرياضيه

-0.55

StoryboardSegue

-0.55

$_['

-0.55

 onOptions

-0.52

 ✭✭

-0.51

ArrowToggle

-0.49

jspb

-0.48

 fédé

-0.47

POSITIVE LOGITS

while

0.56

readystatechange

0.49

WHILE

0.47

While

0.43

wijl

0.41

 While

0.40

 WHILE

0.40

(;;

0.40

 Пока

0.39

Until

0.39

Activations Density 0.002%

parentheses in the text

while loops

No Comments

No Known Activations