INDEX

Explanations

elements related to language use and communication styles in professional contexts

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

slang, jargon, or uncommon languages

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

focusable

-0.33

ardini

-0.30

vall

-0.29

 kebakaran

-0.29

 oseb

-0.29

 vengan

-0.28

 mantenerse

-0.27

getDoctrine

-0.27

idine

-0.27

keda

-0.27

POSITIVE LOGITS

parsedMessage

0.57

 pinulongan

0.55

 dAtA

0.55

 queſta

0.55

ArgsConstructor

0.54

arakhand

0.53

 FetchType

0.52

hyrchwyd

0.50

speech

0.49

 Taktlose

0.49

Activations Density 0.031%

elements related to language use and communication styles in professional contexts

slang, jargon, or uncommon languages

No Comments

No Known Activations