INDEX

Explanations

phrases related to written notes or messages

oai_token-act-pair · gpt-3.5-turbo

references to notes and note-taking materials

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

rates

-0.70

Interstitial

-0.67

ãĥĩ

-0.65

LOS

-0.61

 Mandal

-0.61

ONSORED

-0.60

INA

-0.59

eatures

-0.59

inates

-0.59

 Federation

-0.58

POSITIVE LOGITS

book

1.14

books

1.10

BOOK

1.10

board

0.85

pad

0.85

downs

0.83

boards

0.83

ably

0.82

ariat

0.81

card

0.81

Activations Density 0.052%

phrases related to written notes or messages

references to notes and note-taking materials

No Comments

No Known Activations