INDEX

Explanations

proper nouns such as names of people or places

oai_token-act-pair · gpt-3.5-turbo

references to specific entities or subjects related to nature and technology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Peninsula

-0.73

 Deer

-0.71

NEY

-0.67

baugh

-0.66

oway

-0.65

bred

-0.64

bre

-0.64

eal

-0.64

Interstitial

-0.63

 Sevent

-0.61

POSITIVE LOGITS

 graphene

1.21

 oxide

1.14

 dioxide

0.93

 nanop

0.84

 filament

0.80

 capacitor

0.79

 yarn

0.78

 OLED

0.76

 fibre

0.76

 alloy

0.75

Activations Density 0.006%

proper nouns such as names of people or places

references to specific entities or subjects related to nature and technology

No Comments

No Known Activations