INDEX

Explanations

programming constructs related to counting and returning values

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

numbers and counts

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

yntaxException

-0.80

SOUNDBITE

-0.58

GraphicsUnit

-0.58

 estekak

-0.57

RectangleBorder

-0.54

LookAnd

-0.54

adaptiveStyles

-0.54



-0.53

matchCondition

-0.53

scriptId

-0.53

POSITIVE LOGITS

 count

0.59

count

0.51

Count

0.43

 Count

0.42

 counting

0.40

 counts

0.38

Estimate

0.38

 OFDb

0.37

 getItemCount

0.36

WaitGroup

0.35

Activations Density 0.018%

programming constructs related to counting and returning values

numbers and counts

No Comments

No Known Activations