INDEX

Explanations

punctuation and numerical sequences

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

remainders and multiples

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

featureID

-0.62

reportWebVitals

-0.52

.*")]

-0.48

numerusform

-0.47

fromCharCode

-0.47

 utafitiHapana

-0.47

ntag

-0.45

UpInside

-0.44

ArgsConstructor

-0.44

SuppressLint

-0.43

POSITIVE LOGITS

 autorytatywna

0.54

 CreateTagHelper

0.45

FDRE

0.39

 الرياضيه

0.39

******/

0.38

Merkmale

0.37

makeText

0.37

GEBURTSDATUM

0.36

 zijne

0.36

今回は

0.36

Activations Density 0.001%

punctuation and numerical sequences

remainders and multiples

No Comments

No Known Activations