INDEX

Explanations

model numbers or codes related to electronic devices

oai_token-act-pair · gpt-3.5-turbo

brand names and product identifiers

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

Pru

-0.78

erer

-0.74

owicz

-0.71

enance

-0.70

ufact

-0.69

sson

-0.69

à¨

-0.69

hood

-0.68

 Dunham

-0.67

yard

-0.66

POSITIVE LOGITS

AX

1.08

xon

1.05

MX

0.98

xus

0.95

rx

0.90

yip

0.89

hift

0.89

MX

0.88

yz

0.88

AX

0.86

Activations Density 0.016%

model numbers or codes related to electronic devices

brand names and product identifiers

No Comments

No Known Activations