INDEX

Explanations

information related to historical figures and events

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

thouse

-0.74

plet

-0.69

ophe

-0.66

jri

-0.66

rame

-0.66

brainer

-0.66

terday

-0.66

 shoppers

-0.64

 PLUS

-0.63

 damned

-0.63

POSITIVE LOGITS

Contents

1.21

Appearances

1.19

Trivia

1.08

References

1.07

tymology

0.99

edit

0.99

âĨĳ

0.97

Notable

0.96

History

0.96

Gallery

0.94

Activations Density 4.407%

information related to historical figures and events

No Comments

No Known Activations