INDEX

Explanations

commands and recipes

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

instructional commands

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

phrases that begin procedural instructions or step headings (step-starters like "In/On/If/First" and the following container or UI element).

oai_token-act-pair · gpt-5-mini Triggered by @vetterc0

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

warted

-0.35

<?,

-0.33

Présentation

-0.33

 endwhile

-0.33

templateUrl

-0.33

lyd

-0.33

processData

-0.33

Slf

-0.33

 meinte

-0.33

 blijkt

-0.32

POSITIVE LOGITS

tagHelperRunner

0.71

OGND

0.56

 hãy

0.55

 please

0.55

MemoryWarning

0.54

出版年

0.53

不妨

0.52

bierz

0.51

keteers

0.51

rungsseite

0.50

Activations Density 0.114%

commands and recipes

instructional commands

phrases that begin procedural instructions or step headings (step-starters like "In/On/If/First" and the following container or UI element).

No Comments

No Known Activations