INDEX

Explanations

terms related to design

oai_token-act-pair · gpt-3.5-turbo

references to design concepts and principles

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Arist

-0.79

essional

-0.75

Ton

-0.71

 Priv

-0.69

Tro

-0.69

nikov

-0.69

 Maid

-0.69

zees

-0.68

 sher

-0.67

 Witnesses

-0.67

POSITIVE LOGITS

 design

1.22

yout

1.09

 designs

1.08

 redesign

1.03

 designer

0.99

 designers

0.92

 designing

0.92

smanship

0.89

 Design

0.81

 layouts

0.78

Activations Density 0.020%

terms related to design

references to design concepts and principles

No Comments

No Known Activations