INDEX

Explanations

instances where the letter "a" is followed by a number

oai_token-act-pair · gpt-3.5-turbo

articles and specific quantifiers or indicators of quantity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ESPN

-0.72

AIDS

-0.72

Hon

-0.71

Enjoy

-0.71

Engine

-0.71

America

-0.69

 outfits

-0.69

 Anaheim

-0.69

 Oscar

-0.68

 Orlando

-0.67

POSITIVE LOGITS

lot

1.07

 predetermined

1.01

 specific

1.00

 particular

1.00

 corresponding

0.99

 single

0.99

 variety

0.98

uras

0.98

 definite

0.97

 reasonable

0.97

Activations Density 0.637%

instances where the letter "a" is followed by a number

articles and specific quantifiers or indicators of quantity

No Comments

No Known Activations