INDEX

Explanations

percentages or numerical figures expressed in a particular format

oai_token-act-pair · gpt-3.5-turbo

bullet points or list items in text

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

othal

-0.82

olyn

-0.71

udic

-0.70

erer

-0.69

ierre

-0.69

 graz

-0.67

aughter

-0.64

uve

-0.63

enthal

-0.61

aults

-0.59

POSITIVE LOGITS

Â·Â·

1.23

Â·

0.87

âĢ¢âĢ¢

0.86

¼

0.82

¾

0.76

lat

0.75

µ

0.75

 Joined

0.70

ilities

0.69

ties

0.68

Activations Density 0.013%

percentages or numerical figures expressed in a particular format

bullet points or list items in text

No Comments

No Known Activations