INDEX

Explanations

phrases related to locations or communities

oai_token-act-pair · gpt-3.5-turbo

references to specific groups of people or populations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

lethal

-0.76

 certs

-0.73

wcsstore

-0.70

ctor

-0.64

/"

-0.64

buster

-0.63

Enlarge

-0.63

dayName

-0.62

0000000000000000

-0.62

rang

-0.62

POSITIVE LOGITS

 Alexandria

0.88

 Estonia

0.81

 Qatar

0.81

 Palestine

0.80

 England

0.80

 Crimea

0.79

 southwestern

0.78

 Philadelphia

0.78

 northeastern

0.78

 Toronto

0.78

Activations Density 0.101%

phrases related to locations or communities

references to specific groups of people or populations

No Comments

No Known Activations