INDEX

Explanations

phrases related to the act of writing

oai_token-act-pair · gpt-3.5-turbo

references to various forms of writing and writers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Mund

-0.68

 officials

-0.63

uder

-0.62

 Depot

-0.62

 Last

-0.60

 rear

-0.60

 equip

-0.59

 mass

-0.57

 waist

-0.56

itar

-0.56

POSITIVE LOGITS

writing

4.46

writers

2.77

writer

2.25

written

2.13

write

2.04

writ

1.87

wrote

1.73

Writing

1.68

 writing

1.65

 Writing

1.51

Activations Density 0.015%

phrases related to the act of writing

references to various forms of writing and writers

No Comments

No Known Activations