INDEX

Explanations

numbers and technical terms in a specific format, interspersed with other text

oai_token-act-pair · gpt-3.5-turbo

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

 tremend

-0.76

 hiber

-0.71

ĸļ

-0.70

 Leilan

-0.64

 hoard

-0.64

minist

-0.63

 citiz

-0.63

ometimes

-0.62

 ageing

-0.62

 footing

-0.62

POSITIVE LOGITS

Avg

0.79

Correct

0.77

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

0.74

Total

0.74

Adjust

0.72

Lens

0.71

Reward

0.71

assembly

0.71

total

0.71

ãĥ©ãĥ³

0.70

Activations Density 0.093%

numbers and technical terms in a specific format, interspersed with other text

No Comments

No Known Activations