INDEX

Explanations

references to inventions and their inventors

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

Word follows a verb

np_acts-logits-general · gemini-2.0-flash

invented, support, do, Implement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 iſt

-1.27

 Monfieur

-1.24

 ―――――

-1.24

 itſelf

-1.20

 Jefus

-1.20

 myſelf

-1.20

Efq

-1.18

 againſt

-1.18

 pleaſure

-1.18

 Anſ

-1.14

POSITIVE LOGITS

 multi

0.88

0.87

0.83

 double

0.77

 auto

0.77

 Multi

0.76

“

0.75

0.73

 Auto

0.69

«

0.68

Activations Density 1.778%

references to inventions and their inventors

Word follows a verb

invented, support, do, Implement

No Comments

No Known Activations