INDEX

Explanations

phrases indicating the absence or non-presentation of information

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

technical language about data processing or computational tasks.

oai_token-act-pair · claude-3-5-haiku-20241022 Triggered by @neilrathi

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

ValueStyle

-0.66

Rujuakan

-0.62

HasAnnotation

-0.59

BSITE

-0.56

 виправивши

-0.56

BeginInit

-0.55

ViewImports

-0.53

 Вікі

-0.51

AndEndTag

-0.50

IonicModule

-0.49

POSITIVE LOGITS

 anymore

0.72

 necessarily

0.71

 necesariamente

0.56

nor

0.56

necessarily

0.49

 siquiera

0.48

 lateinit

0.47

 enää

0.47

 forcément

0.47

any

0.46

Activations Density 0.535%

phrases indicating the absence or non-presentation of information

technical language about data processing or computational tasks.

No Comments

No Known Activations