INDEX

Explanations

references to weights or measurements

oai_token-act-pair · gpt-3.5-turbo

references to weight and the act of weighing

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ãĤ¯

-0.70

 Opportun

-0.69

 impunity

-0.65

 Contin

-0.64

uve

-0.62

 Static

-0.61

iji

-0.61

[|

-0.61

 continuity

-0.60

ually

-0.60

POSITIVE LOGITS

 weigh

3.67

 weighed

3.45

 weighing

3.41

 weighs

3.07

 weight

1.67

 weights

1.63

ighed

1.52

 Weight

1.51

Weight

1.46

weight

1.34

Activations Density 0.025%

references to weights or measurements

references to weight and the act of weighing

No Comments

No Known Activations