INDEX

Explanations

terms and structures commonly found in mathematical or formal academic writing

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

mathematical/logical section labels

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

GIVEREF

-0.42

🔙

-0.42

GOTREF

-0.41

 darkness

-0.41

 quedado

-0.40

 vidare

-0.40

 beraber

-0.40

ükemmel

-0.39

lepiej

-0.38

 dolayı

-0.38

POSITIVE LOGITS

 ProtoMessage

0.55

 cherchés

0.54

 gynhyrchwyd

0.52

Rhestr

0.48

expandindo

0.47

 autorytatywna

0.46

 Numerade

0.45

 esternos

0.43

ReadByte

0.43

 وتسجيلات

0.42

Activations Density 0.008%

terms and structures commonly found in mathematical or formal academic writing

mathematical/logical section labels

No Comments

No Known Activations