INDEX

Explanations

structured data representations, particularly those related to code or data entries formatted in a consistent manner

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

`:` followed by code or values

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

исленность

-0.47

 مشين

-0.44

rawDesc

-0.42

AndEndTag

-0.41

 estekak

-0.41

expandindo

-0.40

IsMutable

-0.40

 Wikiseite

-0.40

endpush

-0.39

fromnode

-0.38

POSITIVE LOGITS

 probably

0.47

adaptiveStyles

0.47

Handlung

0.44

 waarschijnlijk

0.43

AccessFile

0.43

吒

0.42

 less

0.42

kenstock

0.42

oubtedly

0.42

yaf

0.42

Activations Density 0.177%

structured data representations, particularly those related to code or data entries formatted in a consistent manner

`:` followed by code or values

No Comments

No Known Activations