INDEX

Explanations

references to things described as "elegant"

oai_token-act-pair · gpt-3.5-turbo

references to elegance and related qualities in descriptions

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

Companies

-0.75

eding

-0.74

VR

-0.72

Downloadha

-0.72

Geo

-0.67

redit

-0.66

Demand

-0.65

Wars

-0.65

 Continued

-0.64

VI

-0.63

POSITIVE LOGITS

 elegance

1.31

 elegant

1.01

 simplicity

0.93

 splend

0.88

isine

0.83

 sophistication

0.82

baum

0.80

 brill

0.80

 masterpiece

0.79

 sophistic

0.78

Activations Density 0.012%

references to things described as "elegant"

references to elegance and related qualities in descriptions

No Comments

No Known Activations