INDEX

Explanations

words related to the mouth and oral activities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Myst

-0.73

 Scotia

-0.68

Syd

-0.66

 Valhalla

-0.65

 Samoa

-0.64

 Anthem

-0.64

 ACPI

-0.64

ARI

-0.63

 Levi

-0.62

 Archdemon

-0.62

POSITIVE LOGITS

piece

1.60

pieces

1.39

parts

1.37

wash

1.35

ful

1.16

feel

1.07

bone

0.97

poke

0.97

pipe

0.95

 watering

0.94

Activations Density 0.039%

words related to the mouth and oral activities

No Comments

No Known Activations