INDEX

Explanations

titles or honorifics associated with people

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

titles

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

 aandacht

-0.34

verifyException

-0.34

 anzu

-0.32

aus

-0.31

Iš

-0.31

seem

-0.30

IBIL

-0.30

原

-0.29

tických

-0.29

 ου

-0.29

POSITIVE LOGITS

parsedMessage

0.74

GTCX

0.66

WEBPACK

0.62

 فريبيس

0.61

 &___

0.59

AxisAlignment

0.59

awtextra

0.57

Cyfarwyddwr

0.56

 otomatig

0.56

 Administrativna

0.55

Activations Density 0.004%

titles or honorifics associated with people

titles

No Comments

No Known Activations