INDEX

Explanations

terms related to burglary and theft

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

burglary

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

 viewDidLoad

-0.59

Personendaten

-0.57

 MutableLiveData

-0.56

oznam

-0.55

 numberWith

-0.54

 Sucher

-0.52

ManagedBean

-0.52

 TextAppearance

-0.52

पया

-0.50

ζ

-0.50

POSITIVE LOGITS

 burglary

0.83

 burglar

0.71

 doors

0.52

 intruder

0.50

 door

0.48

 intruders

0.47

 houſe

0.45

 invasión

0.45

 péné

0.45

 intrusion

0.44

Activations Density 0.009%

terms related to burglary and theft

burglary

No Comments

No Known Activations