INDEX

Explanations

positive and negative opinions or critiques

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

DragonMagazine

-0.99

":""},{"

-0.86

ð

-0.81

idth

-0.76

urrent

-0.74

threat

-0.73

orsi

-0.70

alties

-0.69

OPLE

-0.69

 Newsp

-0.68

POSITIVE LOGITS

 worth

1.11

 enjoyable

1.11

 definitely

1.10

 pretty

1.01

 impressive

1.01

 gorgeous

1.01

 nice

0.99

 fairly

0.98

 catchy

0.97

 quite

0.96

Activations Density 0.238%

positive and negative opinions or critiques

No Comments

No Known Activations