INDEX

Explanations

locations based on their postal abbreviations

oai_token-act-pair · gpt-3.5-turbo

occurrences of U.S. state abbreviations or 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

Not in Any Lists

No Comments

Negative Logits

lihood

-0.94

andowski

-0.71

mania

-0.70

ãĤ¨ãĥ«

-0.69

ÄŁ

-0.68

Versions

-0.65

inav

-0.65

drawn

-0.64

hra

-0.62

urnal

-0.62

POSITIVE LOGITS

./

0.68

 Supervisor

0.68

 Pavilion

0.68

 Area

0.65

ARA

0.65

0.64

JR

0.64

USA

0.63

AMI

0.62

 Territory

0.62

Activations Density 0.039%

locations based on their postal abbreviations

occurrences of U.S. state abbreviations or city names

No Comments

No Known Activations