INDEX

Explanations

instances of the abbreviation "AT"

oai_token-act-pair · gpt-3.5-turbo

references to "AT&T" and related content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

reviewed

-0.80

released

-0.80

ãĤ³

-0.72

perse

-0.72

zanne

-0.70

 Notting

-0.70

ãĥ£

-0.69

uyomi

-0.68

leases

-0.67

Cro

-0.67

POSITIVE LOGITS

AT

0.98

tiny

0.87

AT

0.81

ATA

0.80

anas

0.78

ributes

0.78

ARI

0.78

VIDIA

0.76

hens

0.75

abase

0.74

Activations Density 0.008%

instances of the abbreviation "AT"

references to "AT&T" and related content

No Comments

No Known Activations