INDEX

Explanations

references to printed material or documents

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "print" and its variations, particularly in the context of print media or reproduction

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

asio

-0.77

akia

-0.71

azel

-0.71

 Alban

-0.70

 Galile

-0.67

 Ashton

-0.61

 members

-0.60

 Member

-0.60

bil

-0.58

 member

-0.58

POSITIVE LOGITS

print

1.30

prints

1.19

DragonMagazine

1.06

ertodd

1.03

rint

0.97

printed

0.93

hyde

0.85

itiveness

0.84

gments

0.81

 Printing

0.79

Activations Density 0.009%

references to printed material or documents

occurrences of the word "print" and its variations, particularly in the context of print media or reproduction

No Comments

No Known Activations