INDEX

Explanations

names, particularly of individuals

oai_token-act-pair · gpt-3.5-turbo

social media handles or usernames

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Lis

-0.83

 XIII

-0.80

Vi

-0.79

hack

-0.76

isphere

-0.76

iso

-0.74

 Forth

-0.74

HF

-0.73

yrinth

-0.72

okane

-0.72

POSITIVE LOGITS

Brown

2.80

 Brown

2.72

brown

2.08

 brown

1.83

 Browne

1.78

 Browns

1.43

Gray

1.18

Redd

1.17

 Gray

1.14

Yellow

1.08

Activations Density 0.105%

names, particularly of individuals

social media handles or usernames

No Comments

No Known Activations