INDEX

Explanations

references to rankings or titles of greatness across various fields

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

superlatives and 'best' sequences

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

parsedMessage

-0.58

 ब्रेकडाउन

-0.56

 архивлан

-0.53

sterious

-0.52

RectangleBorder

-0.52

 otomatig

-0.52

 <<<<<<<<<<<<<<

-0.49

 IBOutlet

-0.48

featureID

-0.47

aarrggbb

-0.47

POSITIVE LOGITS

 arguably

0.41

 вовсе

0.37

 mondial

0.37

 classics

0.36

 melhor

0.33

mejores

0.32

 Terbaik

0.32

 برتر

0.32

 absent

0.31

достатки

0.31

Activations Density 0.022%

references to rankings or titles of greatness across various fields

superlatives and 'best' sequences

No Comments

No Known Activations