INDEX

Explanations

sequences of numerical and character patterns

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ValueStyle

-0.74

OGND

-0.52

存于互联网档案馆

-0.51

Personendaten

-0.51

تقاوى

-0.50

RTLR

-0.49

 désolés

-0.48

 BRARY

-0.44

󠁿

-0.43

setArguments

-0.42

POSITIVE LOGITS

 AssemblyVersion

0.42

 odpor

0.41

 bileklik

0.39

 rastre

0.38

essentiel

0.38

 gömlek

0.36

 especialistas

0.35

 대한

0.35

 penga

0.34

Activations Density 0.548%

sequences of numerical and character patterns

No Comments

No Known Activations