INDEX

Explanations

semantic evaluations or rankings

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

icing

-0.62

idon

-0.61

ĪĴ

-0.60

ilation

-0.60

gom

-0.60

 Sharing

-0.59

interrupted

-0.59

¬¼

-0.59

osate

-0.58

 flooded

-0.58

POSITIVE LOGITS

 accordingly

1.51

 differently

1.43

 appropriately

1.18

 separately

1.10

 harshly

0.99

 correctly

0.97

 individually

0.93

 wisely

0.89

 evenly

0.88

 apart

0.87

Activations Density 0.471%

semantic evaluations or rankings

No Known Activations