INDEX

Explanations

proper nouns related to news articles

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

 fortun

-0.73

 interf

-0.73

 polyg

-0.71

 decomp

-0.69

 seiz

-0.69

 photoc

-0.69

 pastry

-0.69

 Marketable

-0.69

 shack

-0.68

çīĪ

-0.67

POSITIVE LOGITS

Ļ

1.42

¬

1.38

ħ

1.27

¤

1.26

Ĵ

1.23

ª

1.22

£

1.21

ĸ

1.20

¡

1.20

ı

1.18

Activations Density 1.552%

proper nouns related to news articles

No Comments

No Known Activations