INDEX

Explanations

mentions of physical seats or positions

oai_token-act-pair · gpt-3.5-turbo

references to seating positions or statuses

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

vernment

-0.79

ogle

-0.75

eff

-0.71

ortium

-0.69

mercial

-0.69

ovych

-0.67

aphael

-0.66

ombies

-0.66

artifacts

-0.65

omb

-0.64

POSITIVE LOGITS

 seat

1.40

seat

1.37

 Seat

1.32

 seats

1.11

holders

0.97

belt

0.97

holder

0.89

bars

0.89

mate

0.86

washer

0.85

Activations Density 0.008%

mentions of physical seats or positions

references to seating positions or statuses

No Comments

No Known Activations