INDEX

Explanations

references to people and their roles or titles in a document

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

reporting season

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

 INTERESAR

-0.42

ConstraintMaker

-0.40

 Verstär

-0.38

انيف

-0.36

 sær

-0.34

 possibilities

-0.34

 duel

-0.34

 فريبيس

-0.34

Datuak

-0.33

 maquillaje

-0.33

POSITIVE LOGITS

艋

0.51

parsedMessage

0.49

 Paglinawan

0.49

Guido

0.48

RemoteException

0.46

 translateY

0.45

enderror

0.45

 comuniques

0.45

IndentedString

0.44

cademic

0.42

Activations Density 0.016%

references to people and their roles or titles in a document

reporting season

No Comments

No Known Activations