INDEX

Explanations

locations or place names

oai_token-act-pair · gpt-3.5-turbo

references to specific locations or places

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

imester

-1.00

ramid

-0.96

aundering

-0.84

okemon

-0.84

uador

-0.81

saf

-0.78

vana

-0.78

inventory

-0.78

ivot

-0.76

zi

-0.76

POSITIVE LOGITS

 Worcester

1.24

cester

0.98

 Middles

0.91

 Lowell

0.87

 Telegram

0.86

 County

0.81

 Somerset

0.80

 Massachusetts

0.80

 Mellon

0.80

 Coliseum

0.78

Activations Density 0.013%

locations or place names

references to specific locations or places

No Comments

No Known Activations