INDEX

Explanations

references to the size or length of data structures or collections

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

length in code

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

enterOuterAlt

-0.68

WriteTagHelper

-0.65

awtextra

-0.65

 referrerpolicy

-0.63

transQ

-0.62

发表于

-0.59

SequentialGroup

-0.59

 BoxFit

-0.57

labelledby

-0.56

MemoryWarning

-0.56

POSITIVE LOGITS

 olev

0.41

 Estadual

0.37

hidupan

0.36

 Económica

0.36

 length

0.35

 Longueur

0.35

getResources

0.35

 Púb

0.35

 Keuangan

0.35

 Panjang

0.34

Activations Density 0.002%

references to the size or length of data structures or collections

length in code

No Comments

No Known Activations