INDEX

Explanations

mentions of political figures or administrations

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

asma

-0.78

Cth

-0.76

 ILCS

-0.76

TOR

-0.76

auri

-0.74

 Edinburgh

-0.72

âĵĺ

-0.71

itionally

-0.71

ounge

-0.70

 Newcastle

-0.67

POSITIVE LOGITS

 administration

0.99

Care

0.98

 appoint

0.95

 Administration

0.90

Obama

0.89

omics

0.87

 Obama

0.86

 vetoed

0.81

 nominated

0.80

 reelection

0.80

Activations Density 0.477%

mentions of political figures or administrations

No Comments

No Known Activations