INDEX

Explanations

the conjunction "and" used in various contexts throughout the document

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

lists joined by and/or

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

Straße

-0.31

tagHelperRunner

-0.30

Quellen

-0.28

 littéraire

-0.28

 miracle

-0.28

yet

-0.28

下去

-0.28

Vielen

-0.27

Allgemeine

-0.27

 mêmes

-0.27

POSITIVE LOGITS

delwed

0.63

 nahilalakip

0.63

StructEnd

0.62

WebVitals

0.59

 AssemblyTitle

0.59

Біографія

0.58

 informée

0.58

rungsseite

0.58

DeleteBehavior

0.57

ьажоргаш

0.56

Activations Density 0.053%

the conjunction "and" used in various contexts throughout the document

lists joined by and/or

No Comments

No Known Activations