INDEX

Explanations

references to specific entities or topics

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

refers to disambiguation

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

новништво

-0.43

alpine

-0.37

nesting

-0.35

 coups

-0.34

ścio

-0.33

Ine

-0.32

 Insured

-0.32

occa

-0.31

 Herbs

-0.31

 Absent

-0.31

POSITIVE LOGITS

 disambiguazione

1.26

verwijspagina

1.05

 ujednoznacz

1.02

曖昧さ回避

0.93

rungsseite

0.85

homonymie

0.80

Personendaten

0.74

 &___

0.70

VYMaps

0.69

ambiguation

0.67

Activations Density 0.008%

references to specific entities or topics

refers to disambiguation

No Comments

No Known Activations

references to specific entities or topics

refers to disambiguation

No Comments

No Known Activations