INDEX

Explanations

proper nouns, specifically company names and associated identifiers

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

company names and organizations

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

梨

-0.41

 հղումներ

-0.39

 sometimes

-0.37

 circumstantial

-0.37

qtype

-0.36

快

-0.36

ři

-0.33

Twitter

-0.33

ingdon

-0.33

得多

-0.33

POSITIVE LOGITS

rungsseite

0.71

ViewImports

0.62

новништво

0.54

 company

0.53

setupUi

0.51

TestBed

0.50

GenerationType

0.50

SQLiteDatabase

0.50

最快更新

0.49

randomUUID

0.48

Activations Density 0.012%

proper nouns, specifically company names and associated identifiers

company names and organizations

No Comments

No Known Activations