INDEX

Explanations

references to common sense and courtesy

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

common sense

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

lorette

-0.45

LayoutPanel

-0.40

 Region

-0.39

 duquel

-0.39

 Barrios

-0.38

 embark

-0.36

 måske

-0.36

 Begegn

-0.35

 region

-0.35

 ihres

-0.35

POSITIVE LOGITS

 commons

0.80

sense

0.77

 Sense

0.74

GEBURTSDATUM

0.71

Sense

0.71

 sensible

0.70

commons

0.70

 sense

0.68

 здра

0.68

UserScript

0.67

Activations Density 0.003%

references to common sense and courtesy

common sense

No Comments

No Known Activations