INDEX

Explanations

programming keywords and symbols associated with data serialization formats

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

indentation and pretty printing

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

 informée

-0.56

endpush

-0.45

 незавершена

-0.44

rxjs

-0.42

zit

-0.41

paramMap

-0.39

RTHOOK

-0.38

entityManager

-0.38

 decided

-0.38

akhstan

-0.38

POSITIVE LOGITS

ModelAdmin

0.46

IntoConstraints

0.45

sidemargin

0.45

featureID

0.43

tonsoft

0.42

WireFormatLite

0.42

xymatrix

0.39

 configureStore

0.38

straßen

0.38

IndentedString

0.38

Activations Density 0.002%

programming keywords and symbols associated with data serialization formats

indentation and pretty printing

No Comments

No Known Activations