INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

proper names, particularly those related to individuals and notable figures

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

ilater

-0.90

ulators

-0.81

urities

-0.78

 indo

-0.76

ulates

-0.73

ular

-0.67

imony

-0.67

ulatory

-0.66

ifier

-0.65

ebus

-0.65

POSITIVE LOGITS

 Doyle

1.32

oyle

1.19

idge

0.87

hyde

0.85

hiba

0.85

ragon

0.78

weed

0.77

mount

0.74

brush

0.74

gaard

0.73

Activations Density 0.008%

proper nouns

proper names, particularly those related to individuals and notable figures

No Comments

No Known Activations