INDEX

Explanations

locations or places

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly places and cities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

cific

-0.79

kefeller

-0.78

ibaba

-0.76

phas

-0.75

antha

-0.69

icing

-0.69

cial

-0.68

hran

-0.66

ificantly

-0.65

iced

-0.65

POSITIVE LOGITS

 COUNTY

1.69

VILLE

1.49

 CITY

1.43

 PARK

1.42

BUR

1.27

 STATE

1.23

FIELD

1.21

 UNIVERS

1.20

SPR

1.19

 SCHOOL

1.17

Activations Density 0.076%

locations or places

proper nouns, particularly places and cities

No Comments

No Known Activations