INDEX

Explanations

mentions of illegal or forbidden activities or items

oai_token-act-pair · gpt-3.5-turbo

mentions of illegal activities or substances

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

soType

-0.85

frames

-0.77

osaurs

-0.77

TOP

-0.76

din

-0.76

ILA

-0.72

packs

-0.72

udeb

-0.71

akeru

-0.71

RM

-0.71

POSITIVE LOGITS

 trafficking

0.97

ly

0.88

 illicit

0.83

 substances

0.83

 goods

0.81

lic

0.81

 enterprises

0.78

 undertaking

0.78

ishly

0.76

 smuggling

0.76

Activations Density 0.015%

mentions of illegal or forbidden activities or items

mentions of illegal activities or substances

No Comments

No Known Activations