INDEX

Explanations

sequences of code related to JSON formatting and output

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_18/width_65k/average_l0_88

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.18.hook_mlp_out

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

曖昧さ回避

-0.68

 InputDecoration

-0.60

 cherchez

-0.56

Rhestr

-0.55

#+#

-0.55

 motherhood

-0.50

 colectiva

-0.50

formatie

-0.50

 polymorphism

-0.49

 graduation

-0.49

POSITIVE LOGITS

Personensuche

0.65

typeorm

0.59

XmlAccessorType

0.57

MemoryWarning

0.54

DeleteBehavior

0.53

ametros

0.53

entait

0.53

++

0.52

 consulté

0.49

ształ

0.49

Activations Density 0.014%

sequences of code related to JSON formatting and output

No Comments

No Known Activations