INDEX

Explanations

terms related to artificiality 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

{\

-0.51

 pribadi

-0.49

ಂತ

-0.48

usiai

-0.47

 înal

-0.47

 oamen

-0.46

 BoxFit

-0.46

usercontent

-0.46

tamment

-0.46

reactivex

-0.45

POSITIVE LOGITS

LookAnd

0.90

 Fake

0.80

 fake

0.73

Fake

0.71

 pretending

0.71

 sembl

0.71

 pretence

0.70

 pretends

0.69

 Mock

0.69

 pretend

0.69

Activations Density 0.225%

terms related to artificiality and deception

No Comments

No Known Activations