INDEX

Explanations

phrases indicating communication or seeking information

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

heard from someone

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

 kasarigan

-0.99

Personendaten

-0.93

 betweenstory

-0.92

expandindo

-0.90

 Мексичка

-0.90

 kaarangay

-0.89

 الحره

-0.87

 Paglinawan

-0.87

 ویکی‌پدیا

-0.87

 AspNetCore

-0.84

POSITIVE LOGITS

 heard

1.30

 hear

1.25

 hearing

1.07

 hears

1.02

Heard

0.97

hear

0.93

heard

0.93

 Heard

0.91

Hear

0.88

Hearing

0.83

Activations Density 0.048%

phrases indicating communication or seeking information

heard from someone

No Comments

No Known Activations