INDEX

Explanations

locations or places mentioned in 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_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

VIDIA

-0.82

gey

-0.75

nikov

-0.70

alys

-0.69

 certs

-0.68

HTML

-0.68

verbal

-0.68

èĢħ

-0.67

 appoint

-0.67

POSITIVE LOGITS

 southwestern

1.30

 northeastern

1.26

 northern

1.24

 southeastern

1.24

 streets

1.23

 downtown

1.22

 southern

1.18

 northwestern

1.18

 nearby

1.17

 Northeast

1.16

Activations Density 2.320%

locations or places mentioned in news articles

No Comments

No Known Activations