INDEX

Explanations

terms related to support and approval, such as endorsements and acknowledgments

oai_token-act-pair · gpt-3.5-turbo

references to endorsements and acknowledgments related to opinions or statements

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

stant

-0.69

ordinate

-0.69

illac

-0.68

ovy

-0.68

hed

-0.67

ebus

-0.67

resp

-0.67

ovych

-0.66

esthetic

-0.66

mers

-0.65

POSITIVE LOGITS

 thereof

0.83

DragonMagazine

0.81

 acknowledgement

0.76

xual

0.74

ism

0.73

naire

0.72

anguage

0.72

ary

0.70

eering

0.70

ãĥĭ

0.70

Activations Density 0.094%

terms related to support and approval, such as endorsements and acknowledgments

references to endorsements and acknowledgments related to opinions or statements

No Comments

No Known Activations