INDEX

Explanations

words related to long-term patterns or changes

oai_token-act-pair · gpt-3.5-turbo

references to trends or patterns over time

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ned

-0.83

oÄŁ

-0.83

gha

-0.79

ded

-0.75

INGTON

-0.71

\\\\\\\\

-0.68

unts

-0.67

lain

-0.66

×ŀ

-0.65

\\\\\\\\\\\\\\\\

-0.65

POSITIVE LOGITS

etting

1.06

ettings

1.03

etter

1.01

uggest

0.96

omething

0.94

afety

0.92

ynt

0.92

 trends

0.91

hooting

0.90

hips

0.89

Activations Density 0.034%

words related to long-term patterns or changes

references to trends or patterns over time

No Comments

No Known Activations