INDEX

Explanations

numerical representations and their format variations

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

hexadecimal character codes

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

expandindo

-0.74

IndentedString

-0.70

InjectAttribute

-0.61

 كومونز

-0.61

AndroidJUnit

-0.55

rungsseite

-0.55

 OMITBAD

-0.53

﹧

-0.50

IsMutable

-0.49

MockBean

-0.49

POSITIVE LOGITS

 vermelhas

0.47

 contenuti

0.42

lò

0.40

 betweenstory

0.40

 visualización

0.39

 legais

0.39

 Herkunft

0.38

 egentligen

0.37

 terecht

0.36

 TAMBÉM

0.36

Activations Density 0.118%

numerical representations and their format variations

hexadecimal character codes

No Comments

No Known Activations