INDEX

Explanations

network-related terms and operations

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

cpl_ message

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

setVerticalGroup

-0.65

 للمعارف

-0.62

 useAppContext

-0.61

 useStyles

-0.56

IndentedString

-0.56

fjspx

-0.55

WriteBarrier

-0.54

ioutil

-0.54

__(/*!

-0.54

URLException

-0.53

POSITIVE LOGITS

 Nachricht

0.33

…………………………………………

0.32

 napraw

0.31

 повідом

0.31

Ծանոթ

0.31

 Lungen

0.30

 message

0.30

 Botschaft

0.30

kyl

0.29

ส่ง

0.28

Activations Density 0.189%

network-related terms and operations

cpl_ message

No Comments

No Known Activations