INDEX

Explanations

unique identifiers or IDs within a data structure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 anything

-0.49

 wszyst

-0.47

 surla

-0.47

 everything

-0.47

 practically

-0.45

 allgemeinen

-0.45

pushFollow

-0.44

 complementary

-0.43

XmlAccessType

-0.42

 Präsidenten

-0.42

POSITIVE LOGITS

 estekak

0.69

 nakalista

0.54

AndEndTag

0.53

 disambiguazione

0.51

parsedMessage

0.50

GEBURTSDATUM

0.48

HostException

0.47

 Roskov

0.46

BufferException

0.46

 виправивши

0.46

Activations Density 0.163%

unique identifiers or IDs within a data structure

No Comments

No Known Activations