INDEX

Explanations

phrases related to close relationships or family members

oai_token-act-pair · gpt-3.5-turbo

references to loved ones

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

illin

-0.87

soDeliveryDate

-0.78

interstitial

-0.74

authorized

-0.74

SPONSORED

-0.73

manship

-0.73

ional

-0.71

arat

-0.70

akedown

-0.68

arb

-0.67

POSITIVE LOGITS

 uncond

0.79

 dearly

0.77

 sacrific

0.70

 Ones

0.70

lihood

0.68

joy

0.67

tsky

0.66

 Heavenly

0.65

 loved

0.64

 Serve

0.63

Activations Density 0.030%

phrases related to close relationships or family members

references to loved ones

No Comments

No Known Activations