INDEX

Explanations

terms and concepts related to fraudulent activities, including various forms of fraud and deception

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

TagMode

-0.75

gypti

-0.64

 surla

-0.61

Źródło

-0.60

fromnode

-0.58

లాలు

-0.58

WireFormatLite

-0.57

ogaster

-0.56

 BoxFit

-0.56

YLON

-0.56

POSITIVE LOGITS

 perpetrated

0.85

 committed

0.66

 honte

0.64

onAttach

0.63

 Prevention

0.61

 danni

0.59

 perpetuated

0.59

 Detection

0.59

 prevention

0.58

 punishable

0.57

Activations Density 0.431%

terms and concepts related to fraudulent activities, including various forms of fraud and deception

No Comments

No Known Activations