INDEX

Explanations

student names

oai_token-act-pair · gpt-3.5-turbo

tokens that represent identifiers or potential data points

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

 Stew

-0.80

 disg

-0.72

 wolves

-0.68

 seiz

-0.67

 eleph

-0.67

 Mall

-0.67

 microw

-0.66

 Insp

-0.66

 Sally

-0.64

POSITIVE LOGITS

1.19

obar

1.00

Bib

0.97

bis

0.97

bs

0.95

bish

0.95

bar

0.94

bin

0.92

baugh

0.91

Activations Density 0.178%

student names

tokens that represent identifiers or potential data points

No Comments

No Known Activations