INDEX

Explanations

mentions of tying, wrapping, or fastening something

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

 Booker

-0.76

vironments

-0.75

 Atmosp

-0.73

renheit

-0.72

geist

-0.72

 Auditor

-0.71

illery

-0.71

elsius

-0.70

Tax

-0.70

PokÃ©

-0.70

POSITIVE LOGITS

 knot

1.63

 rope

1.61

 ropes

1.59

 cords

1.57

 knots

1.57

 straps

1.50

 tether

1.32

 elastic

1.31

 cord

1.31

tie

1.31

Activations Density 0.302%

mentions of tying, wrapping, or fastening something

No Comments

No Known Activations