INDEX

Explanations

references to criminal history and offenses

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

criminal history habitual offender

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

IContainer

-0.60

ArgumentParser

-0.50

expandindo

-0.49

providedIn

-0.49

XmlAccessType

-0.48

ьаж

-0.47

aini

-0.47

SpringBootTest

-0.46

interopRequire

-0.46

󠁬

-0.45

POSITIVE LOGITS

 Normdatei

0.51

 criminal

0.43

 felony

0.42

CppMethod

0.40

 habet

0.38

 disambiguazione

0.37

 convictions

0.36

withIdentifier

0.36

 Criminal

0.35

 mapStateToProps

0.35

Activations Density 0.037%

references to criminal history and offenses

criminal history habitual offender

No Comments

No Known Activations