INDEX

Explanations

phrases related to expertise and knowledge

oai_token-act-pair · gpt-3.5-turbo

negative phrases possibly related to critiques or evaluations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

AVG

-0.69

 Klux

-0.64

Moj

-0.60

 confidence

-0.60

 Saunders

-0.60

ABE

-0.58

 Bohem

-0.58

 mainline

-0.58

Cec

-0.58

 Sierra

-0.57

POSITIVE LOGITS

based

1.04

of

1.01

backed

0.99

filled

0.99

busters

0.96

finding

0.95

eye

0.95

tale

0.94

taker

0.94

laden

0.93

Activations Density 0.101%

phrases related to expertise and knowledge

negative phrases possibly related to critiques or evaluations

No Comments

No Known Activations