INDEX

Explanations

punctuation marks and structural elements in code or text

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

user turns

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

AnchorTagHelper

-0.65

 BoxFit

-0.62

 AssemblyCompany

-0.62

.*")]

-0.59

 betweenstory

-0.58

transQ

-0.57

AlterField

-0.56

HomeAsUpEnabled

-0.56

Personendaten

-0.52

 useAppContext

-0.51

POSITIVE LOGITS

SequentialGroup

0.53

Thiết

0.42

Rè

0.40

 utafiti

0.39

niająca

0.38

RefNanny

0.38

Искәрмәләр

0.37

 ***!

0.36

Diweddarwch

0.36

 validamos

0.35

Activations Density 0.000%

punctuation marks and structural elements in code or text

user turns

No Comments

No Known Activations