INDEX

Explanations

phrases expressing openness and acceptance

oai_token-act-pair · gpt-3.5-turbo

instances of the word "welcome" indicating acceptance or approval

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

iph

-0.86

iple

-0.79

olester

-0.75

hid

-0.74

orp

-0.73

negie

-0.72

thin

-0.69

pel

-0.67

arcity

-0.66

oret

-0.65

POSITIVE LOGITS

 welcome

1.30

elcome

1.15

 welcomes

1.06

 Welcome

0.93

Welcome

0.91

 welcomed

0.87

 welcoming

0.82

ãĤĵ

0.80

 laughter

0.80

 applause

0.80

Activations Density 0.011%

phrases expressing openness and acceptance

instances of the word "welcome" indicating acceptance or approval

No Comments

No Known Activations