INDEX

Explanations

mathematical expressions and structural components in technical documents

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

terrace C atoms

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

Personensuche

-0.52

writeFieldEnd

-0.52

queryInterface

-0.50

 dentaire

-0.49

पया

-0.48

ItemBackground

-0.47

astéroïdes

-0.45

 autorytatywna

-0.45

 OMITBAD

-0.44

出版年

-0.44

POSITIVE LOGITS

 journeys

0.40

 longer

0.40

 unknownFields

0.38

 Routine

0.38

lox

0.38

 drift

0.37

 patrols

0.37

seiten

0.37

 interior

0.36

 routine

0.36

Activations Density 0.274%

mathematical expressions and structural components in technical documents

terrace C atoms

No Comments

No Known Activations