INDEX

Explanations

references to programming constructs and entities such as classes, methods, and data types

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 surla

-0.54

transQ

-0.51

tori

-0.46

IContainer

-0.46

 PROV

-0.46

Lugares

-0.46

 castor

-0.45

ipto

-0.44

SIT

-0.44

енча

-0.44

POSITIVE LOGITS

Spécification

0.48

paravant

0.43

 logiciel

0.42

 judiciales

0.41

GOTREF

0.41

internalType

0.41

 незавершена

0.40

 Winaray

0.40

ThroughAttribute

0.40

 république

0.39

Activations Density 0.044%

references to programming constructs and entities such as classes, methods, and data types

No Comments

No Known Activations