INDEX

Explanations

text within quotation marks

oai_token-act-pair · gpt-3.5-turbo

quotation marks and their contents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 surpr

-0.73

 plateau

-0.72

 dogged

-0.71

etheless

-0.70

 moderates

-0.68

 tire

-0.67

 Fisheries

-0.67

 radicals

-0.66

 grav

-0.65

 dipping

-0.62

POSITIVE LOGITS

./

1.18

SELECT

1.13

BuyableInstoreAndOnline

1.08

../

1.02

Hello

1.00

wcsstore

0.93

Usage

0.93

foo

0.91

default

0.90

<?

0.88

Activations Density 0.047%

text within quotation marks

quotation marks and their contents

No Comments

No Known Activations