INDEX

Explanations

references to specific names, particularly notable individuals

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

Rachel and Watson individuals

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

enumi

-0.40

In

-0.35

yii

-0.34

by

-0.34

 Currie

-0.33

-0.32

 Herv

-0.32

 Harwood

-0.32

/,

-0.32

際は

-0.32

POSITIVE LOGITS

Rachel

0.91

ACHEL

0.89

 Rachel

0.87

Zeneca

0.80

rachel

0.78

Watson

0.75

achel

0.73

 WATSON

0.72

 Rashford

0.70

 Watson

0.69

Activations Density 0.024%

references to specific names, particularly notable individuals

Rachel and Watson individuals

No Comments

No Known Activations