INDEX

Explanations

names containing "Eli" and "Lilly"

oai_token-act-pair · gpt-3.5-turbo

mentions of specific names, particularly notable individuals and characters

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

rav

-0.79

 Moroccan

-0.79

interrupted

-0.74

assault

-0.72

 Commonwealth

-0.72

 Chambers

-0.71

 Speedway

-0.71

 Sabha

-0.70

undy

-0.70

colour

-0.68

POSITIVE LOGITS

Eli

2.62

Ot

2.30

hei

2.03

ot

2.01

 Nico

1.39

Shu

1.37

Kot

1.35

 Mari

1.23

 Levi

1.19

Aki

1.18

Activations Density 0.049%

names containing "Eli" and "Lilly"

mentions of specific names, particularly notable individuals and characters

No Comments

No Known Activations