INDEX

Explanations

references to geographical locations, specifically continents and countries

oai_token-act-pair · gpt-3.5-turbo

references to geographic locations, specifically continents and mainland areas

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

HUD

-0.81

Room

-0.79

PAC

-0.78

CHAT

-0.73

Roy

-0.72

oga

-0.71

Focus

-0.70

lett

-0.69

endo

-0.69

adi

-0.69

POSITIVE LOGITS

 Continent

0.85

 continents

0.81

 continental

0.80

continental

0.79

 drift

0.79

 shelf

0.78

 continent

0.76

 mainland

0.75

 shelves

0.73

uania

0.73

Activations Density 0.033%

references to geographical locations, specifically continents and countries

references to geographic locations, specifically continents and mainland areas

No Comments

No Known Activations