INDEX

Explanations

image captions with credits to photo sources

oai_token-act-pair · gpt-3.5-turbo

references to images and their sources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 bicycl

-0.78

 restroom

-0.74

 honorable

-0.74

 advis

-0.73

 statewide

-0.73

 honored

-0.70

leans

-0.68

 centers

-0.68

 elevator

-0.68

 honors

-0.68

POSITIVE LOGITS

Â£

1.01

 Â£

0.99

 Photograph

0.99

BBC

0.98

 ÂŃ

0.97

 envis

0.92

UK

0.90

Scotland

0.90

 Nato

0.89

Britain

0.87

Activations Density 0.054%

image captions with credits to photo sources

references to images and their sources

No Comments

No Known Activations