INDEX

Explanations

words related to items or activities commonly found on a desk

oai_token-act-pair · gpt-3.5-turbo

references to desks

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

Interstitial

-0.81

Ô

-0.80

 Adin

-0.76

Norm

-0.70

-0.69

66666666

-0.67

ETA

-0.67

UGC

-0.65

-0.64

OUND

-0.63

POSITIVE LOGITS

tops

1.50

 desk

1.30

 drawer

1.19

 desks

0.98

 Desk

0.97

wu

0.90

chair

0.88

 chair

0.88

 clerk

0.88

 attendant

0.86

Activations Density 0.006%

words related to items or activities commonly found on a desk

references to desks

No Comments

No Known Activations