INDEX

Explanations

locations related to different cities

oai_token-act-pair · gpt-3.5-turbo

mentions of "San" as part of city names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

•San

#1 activation: 85.53

•Fr

#4 activation: 9.32

No Comments

Negative Logits

ï¸ı

-0.66

ICS

-0.66

 numbering

-0.64

 prompt

-0.63

 numbered

-0.60

wig

-0.60

 wise

-0.58

 markup

-0.57

puff

-0.57

 subscribing

-0.57

POSITIVE LOGITS

 Francisco

1.41

 Diego

1.33

ctuary

1.26

 Bernardino

1.18

itary

1.17

 Antonio

1.12

San

1.09

Franc

1.07

Disk

1.06

ibel

1.05

Activations Density 0.024%

locations related to different cities

mentions of "San" as part of city names

No Comments

No Known Activations