INDEX

Explanations

numeric values related to sizes or areas

oai_token-act-pair · gpt-3.5-turbo

references to square footage measurements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

eer

-0.74

ARP

-0.71

oan

-0.68

 Claus

-0.67

 Honour

-0.66

giving

-0.66

 appropri

-0.63

 Barcl

-0.63

 Democrat

-0.62

vironment

-0.62

POSITIVE LOGITS

pees

0.98

 Enix

0.91

feet

0.87

µ

0.87

 kilometres

0.79

 Feet

0.78

 kilomet

0.78

square

0.77

sq

0.76

 kilometers

0.76

Activations Density 0.010%

numeric values related to sizes or areas

references to square footage measurements

No Comments

No Known Activations