INDEX

Explanations

numerical measurements related to weight

oai_token-act-pair · gpt-3.5-turbo

numerical weights and measurements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

otiation

-0.87

 Opportun

-0.76

nesday

-0.74

 backdoor

-0.72

atre

-0.71

timeout

-0.68

ctuary

-0.67

Gate

-0.67

iphate

-0.66

acular

-0.66

POSITIVE LOGITS

 weighs

1.34

 weighing

1.30

 kilograms

1.26

 weigh

1.23

lbs

1.23

 pounds

1.23

 weights

1.21

 weight

1.20

lbs

1.19

kg

1.19

Activations Density 0.173%

numerical measurements related to weight

numerical weights and measurements

No Comments

No Known Activations