INDEX

Explanations

mentions of the concept of something being created or comprised of various components

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

assis

-0.58

CLOSE

-0.57

ities

-0.56

hov

-0.55

idon

-0.54

CPC

-0.53

iveness

-0.53

GPA

-0.52

Cul

-0.52

isks

-0.52

POSITIVE LOGITS

 payable

1.09

 available

0.92

 redundant

0.88

 obsolete

0.87

 possible

0.85

 famous

0.85

 aware

0.84

usable

0.81

lyn

0.80

 landfall

0.78

Activations Density 0.061%

mentions of the concept of something being created or comprised of various components

No Comments

No Known Activations