INDEX

Explanations

notes or annotations within a text

oai_token-act-pair · gpt-3.5-turbo

instances of notes or annotations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

asting

-0.72

heast

-0.72

 hair

-0.69

 lodging

-0.69

 school

-0.69

 waged

-0.68

 volunte

-0.68

 freight

-0.67

 revital

-0.66

esc

-0.66

POSITIVE LOGITS

 Note

1.11

 Notes

1.06

BOOK

1.02

books

1.00

book

0.93

 note

0.92

 NOTE

0.90

Note

0.88

Notes

0.86

Pad

0.83

Activations Density 0.012%

notes or annotations within a text

instances of notes or annotations

No Comments

No Known Activations