INDEX

Explanations

addresses and locations

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

Data Type

torch.float32

Hook Point

blocks.8.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

erity

-0.72

TPP

-0.72

rators

-0.69

icators

-0.67

 proteins

-0.67

 negotiator

-0.64

BILITIES

-0.63

ubi

-0.62

xual

-0.62

aper

-0.62

POSITIVE LOGITS

 northwest

1.35

 southwest

1.32

 southwestern

1.28

 suburb

1.25

 northeast

1.25

 southeastern

1.25

 northwestern

1.22

 southeast

1.21

 northeastern

1.17

 east

1.13

Activations Density 4.938%

addresses and locations

No Comments

No Known Activations