INDEX

Explanations

references to the human eye

oai_token-act-pair · gpt-3.5-turbo

references to the word "eye"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Vide

-0.68

emonium

-0.66

minus

-0.65

 Violent

-0.63

 Resistance

-0.63

theless

-0.62

stood

-0.62

channelAvailability

-0.61

 Silicon

-0.59

lishes

-0.59

POSITIVE LOGITS

 sockets

1.10

eyed

0.90

eye

0.89

ight

0.86

bones

0.86

 lobe

0.84

 socket

0.84

 eyebrow

0.83

bone

0.83

 brow

0.82

Activations Density 0.013%

references to the human eye

references to the word "eye"

No Comments

No Known Activations