INDEX

Explanations

news-related entities such as organizations 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_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

gery

-0.73

gie

-0.71

holders

-0.68

dies

-0.68

sed

-0.67

erva

-0.66

wikipedia

-0.66

warm

-0.66

 scrib

-0.66

gart

-0.66

POSITIVE LOGITS

 affiliate

1.02

Money

1.00

 News

0.98

NEWS

0.97

News

0.95

Universal

0.91

DEF

0.91

ILLE

0.85

IJ

0.85

 reporter

0.84

Activations Density 0.415%

news-related entities such as organizations and locations

No Comments

No Known Activations