INDEX

Explanations

URLs and requests for online information

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

"visit" followed by a URL

np_acts-logits-general · gemini-2.0-flash

visit followed by URLs

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

ſelf

-0.68

 ſta

-0.63

 abſ

-0.60

 ModelExpression

-0.58

ſelves

-0.58

 ſte

-0.57

 deſt

-0.57

 ſever

-0.56

 ſol

-0.55

 poffe

-0.55

POSITIVE LOGITS

the

0.41

our

0.40

://

0.38

nė

0.37

 Bright

0.36

Bry

0.34

bright

0.34

ActivityCompat

0.34

putable

0.34

web

0.33

Activations Density 0.187%

URLs and requests for online information

"visit" followed by a URL

visit followed by URLs

No Comments

No Known Activations