INDEX

Explanations

words related to labeling or applying labels

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

issance

-0.72

ctica

-0.70

skill

-0.69

aldo

-0.66

rouch

-0.65

vati

-0.64

 Pradesh

-0.62

atana

-0.61

ashington

-0.60

isites

-0.60

POSITIVE LOGITS

 label

0.87

 labels

0.86

mates

0.83

ovan

0.80

strip

0.80

cloth

0.79

printed

0.77

mate

0.76

 Label

0.76

mark

0.74

Activations Density 0.048%

words related to labeling or applying labels

No Comments

No Known Activations