INDEX

Explanations

numerical data related to time and version information

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

Dates, often with colons or slashes

np_acts-logits-general · gemini-2.0-flash

version numbers and dates

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

IsContent

-0.61

 wikipagina

-0.59

 MonoBehaviour

-0.58

DoubleQuotes

-0.57

SequentialGroup

-0.56

 kağıdı

-0.54

simpleType

-0.52

Hentet

-0.52

Javadoc

-0.51

AnchorStyles

-0.51

POSITIVE LOGITS

 Trainer

0.43

 Thermal

0.37

 sopr

0.37

postsleuth

0.36

 احت

0.35

 betweenstory

0.35

 afic

0.34

Cau

0.33

TUT

0.33

ugc

0.33

Activations Density 0.150%

numerical data related to time and version information

Dates, often with colons or slashes

version numbers and dates

No Comments

No Known Activations