INDEX

Explanations

references to the word "tor" in URLs

oai_token-act-pair · gpt-3.5-turbo

references to the "Tor" technology or network

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

IU

-0.70

 maize

-0.66

 cows

-0.65

 cooperative

-0.65

 GOODMAN

-0.62

Fe

-0.61

agher

-0.60

 friendly

-0.59

 Iowa

-0.59

 Guam

-0.59

POSITIVE LOGITS

tor

4.77

Tor

1.93

TOR

1.77

tion

1.68

tur

1.59

tor

1.58

TOR

1.42

Tor

1.37

tions

1.34

tar

1.21

Activations Density 0.016%

references to the word "tor" in URLs

references to the "Tor" technology or network

No Comments

No Known Activations