INDEX

Explanations

different forms or contexts of the word "representation"

oai_token-act-pair · gpt-3.5-turbo

references to representation concepts

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

launch

-0.75

few

-0.73

cake

-0.71

awar

-0.71

strap

-0.71

imb

-0.70

stead

-0.69

sterdam

-0.66

foot

-0.66

nen

-0.65

POSITIVE LOGITS

 Represent

1.00

ational

0.93

 representation

0.90

 representations

0.88

ative

0.86

ATIVE

0.86

eering

0.85

atively

0.83

DonaldTrump

0.81

represented

0.74

Activations Density 0.023%

different forms or contexts of the word "representation"

references to representation concepts

No Comments

No Known Activations