INDEX

Explanations

the word "not"

oai_token-act-pair · gpt-3.5-turbo

negative phrases or statements indicating what should not be done

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

ELD

-0.75

quickShipAvailable

-0.74

anders

-0.71

 Inventory

-0.70

 Appears

-0.65

itiz

-0.64

lined

-0.63

soDeliveryDate

-0.61

å¥

-0.60

Line

-0.60

POSITIVE LOGITS

 necessarily

1.22

icably

1.21

icable

1.04

 hesitate

0.91

epad

0.89

 bother

0.86

necess

0.85

 exactly

0.84

ember

0.83

 suffice

0.80

Activations Density 0.081%

the word "not"

negative phrases or statements indicating what should not be done

No Comments

No Known Activations