INDEX

Explanations

words related to deception and falsehood

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

orianCalendar

-0.56

iibo

-0.56

ALF

-0.52

bootstrapcdn

-0.52

iettes

-0.50

 ویکی‌پدی

-0.48

HomeAsUpEnabled

-0.48

IFA

-0.47

FID

-0.46

ysuckle

-0.46

POSITIVE LOGITS

 étoient

0.60

 avoient

0.56

 sindic

0.53

 grecque

0.51

 venezolano

0.51

 coffret

0.50

 advogado

0.50

 femininas

0.50

 lienzo

0.49

 négo

0.49

Activations Density 1.278%

words related to deception and falsehood

No Comments

No Known Activations