INDEX

Explanations

phrases related to feeling physically or emotionally comfortable

oai_token-act-pair · gpt-3.5-turbo

expressions of comfort and comfort-related themes

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

oys

-0.73

 Destruction

-0.71

 Strikes

-0.70

 Mole

-0.68

 Fail

-0.68

 archae

-0.65

oids

-0.63

III

-0.63

olan

-0.62

RIP

-0.62

POSITIVE LOGITS

 comfortable

3.51

 uncomfortable

1.92

fortable

1.85

 comfortably

1.81

 cozy

1.62

 comfort

1.52

 confident

1.48

 accustomed

1.38

 snug

1.38

comfort

1.37

Activations Density 0.014%

phrases related to feeling physically or emotionally comfortable

expressions of comfort and comfort-related themes

No Comments

No Known Activations