INDEX

Explanations

phrases related to requests for information or leads in a police investigation

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

wanted by police

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

évaluateur

-0.34

 ModelExpression

-0.33

SizeMode

-0.32

seg

-0.32

 soft

-0.30

webElement

-0.29

 slow

-0.28

SourceChecksum

-0.28

 lead

-0.28

POSITIVE LOGITS

 ainfi

0.70

LookAnd

0.69

 étoit

0.67

 Komunikasi

0.63

majánló

0.61

 Geſ

0.60

 detectives

0.59

 Announcements

0.58

 Verſ

0.58

 Dete

0.57

Activations Density 0.005%

phrases related to requests for information or leads in a police investigation

wanted by police

No Comments

No Known Activations