INDEX

Explanations

elements, attributes, and structural components in data representation and code

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

conviction, crime, or rendering

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

structural, non-content tokens—common function words, punctuation/quotes, and markup or notation symbols that organize text rather than convey topic content.

oai_token-act-pair · gpt-5 Triggered by @vetterc0

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

NameInMap

-0.52

InjectAttribute

-0.50

 Biôgrafia

-0.45

colgante

-0.42

Rujuakan

-0.41

WithIOException

-0.39

aarrggbb

-0.36

Jîn

-0.36

 blé

-0.35

 delantero

-0.35

POSITIVE LOGITS

 يتيمه

0.51

endregion

0.50

 ब्रेकडाउन

0.46

RenderAtEndOf

0.44

 onCreateView

0.42

richTextPanel

0.41

HtmlAttribute

0.41

 oprot

0.40

Portail

0.40

transQ

0.39

Activations Density 0.135%

elements, attributes, and structural components in data representation and code

conviction, crime, or rendering

structural, non-content tokens—common function words, punctuation/quotes, and markup or notation symbols that organize text rather than convey topic content.

No Comments

No Known Activations

elements, attributes, and structural components in data representation and code

conviction, crime, or rendering

structural, non-content tokens—common function words, punctuation/quotes, and markup or notation symbols that organize text rather than convey topic content.

No Comments

No Known Activations