INDEX

Explanations

locations or entities related to the geographical region or concept represented by 'Tier'

oai_token-act-pair · gpt-3.5-turbo

references to specific tiers or levels, particularly in a context that suggests ranking or categorization

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

etsk

-0.91

endi

-0.79

DragonMagazine

-0.76

ga

-0.75

ovych

-0.73

ocrat

-0.72

 dipping

-0.71

 brill

-0.71

ascus

-0.70

ocrats

-0.68

POSITIVE LOGITS

 Tier

2.42

 Triple

1.78

Tier

1.63

 Route

1.62

 Gamble

1.32

Rou

1.28

 Pillar

1.25

Ley

1.20

 Grade

1.19

Pit

1.19

Activations Density 0.060%

locations or entities related to the geographical region or concept represented by 'Tier'

references to specific tiers or levels, particularly in a context that suggests ranking or categorization

No Comments

No Known Activations