INDEX

Explanations

references to specific entities or subjects within a context

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

relative clauses

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

 focusing

-0.34

 emerging

-0.32

 meaning

-0.32

 seeking

-0.32

 señalado

-0.31

low

-0.31

Des

-0.30

 énergétique

-0.30

Re

-0.30

 weighing

-0.30

POSITIVE LOGITS

addContainerGap

0.82

 myſelf

0.70

 itſelf

0.68

 الرياضيه

0.67

 оригіналу

0.60

 Administrativna

0.59

NameInMap

0.58

Tembelea

0.56

 Beſ

0.56

 BorderRadius

0.56

Activations Density 0.022%

references to specific entities or subjects within a context

relative clauses

No Comments

No Known Activations