INDEX

Explanations

references to specific institutions or locations, particularly art museums

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 blocker

-0.88

yip

-0.78

ĪĴ

-0.73

 stride

-0.72

 blockers

-0.72

nexus

-0.69

 accelerator

-0.69

dash

-0.68

brainer

-0.68

ibling

-0.66

POSITIVE LOGITS

 Contemporary

1.06

 Photography

0.97

 Antiqu

0.95

 Hours

0.95

 Archives

0.91

 History

0.90

 Computing

0.89

 Museum

0.87

 Archae

0.86

 Mysteries

0.85

Activations Density 0.082%

references to specific institutions or locations, particularly art museums

No Comments

No Known Activations