INDEX

Explanations

text related to entities, trademarks, or properties

oai_token-act-pair · gpt-3.5-turbo

symbols or punctuation that disrupt the text formatting

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

 scatter

-0.81

 dirt

-0.64

itably

-0.63

 damaged

-0.62

 weighted

-0.61

 detached

-0.61

 staggered

-0.60

 Golem

-0.60

anwhile

-0.59

 white

-0.58

POSITIVE LOGITS

âĢ

1.00

¹

0.97

âĢł

0.94

£

0.90

âĹ¼

0.88

į

0.87

º

0.87

¡

0.83

Â¢

0.81

Â§

0.81

Activations Density 0.822%

text related to entities, trademarks, or properties

symbols or punctuation that disrupt the text formatting

No Comments

No Known Activations