INDEX

Explanations

proper nouns or names

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

yrinth

-0.84

irth

-0.74

elaide

-0.74

istar

-0.73

isoft

-0.72

asio

-0.70

romy

-0.68

psey

-0.68

icult

-0.68

akings

-0.67

POSITIVE LOGITS

plates

1.49

paces

1.33

plate

1.27

names

1.07

paced

1.00

 names

0.94

ames

0.90

 aliases

0.89

calling

0.87

brand

0.85

Activations Density 0.043%

proper nouns or names

No Comments

No Known Activations