INDEX

Explanations

mentions of art-related terms and concepts

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Witnesses

-0.72

iod

-0.69

 Adin

-0.67

 vaccination

-0.67

 Ticket

-0.66

DN

-0.65

 Limbaugh

-0.63

Availability

-0.62

 Spears

-0.61

olate

-0.60

POSITIVE LOGITS

emis

1.20

Artist

1.14

smanship

1.13

isans

1.06

 artwork

1.00

 artist

0.98

artist

0.98

urally

0.96

art

0.95

 arts

0.94

Activations Density 2.597%

mentions of art-related terms and concepts

No Comments

No Known Activations