INDEX

Explanations

references to photos

oai_token-act-pair · gpt-3.5-turbo

references to photographs

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

ensable

-0.84

ressive

-0.78

Mand

-0.78

merce

-0.76

rane

-0.71

cffff

-0.71

ricular

-0.71

resses

-0.69

iveness

-0.69

vernment

-0.67

POSITIVE LOGITS

journal

1.07

photos

1.07

 photo

1.04

 photographed

0.95

photo

0.92

 illustration

0.91

 Phot

0.91

>>>>>>>>

0.89

 phot

0.89

 Photographer

0.89

Activations Density 0.015%

references to photos

references to photographs

No Comments

No Known Activations