INDEX

Explanations

numeric values

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

 Âł

-0.27

 ÂłÂł

-0.25

[/

-0.25

Âł

-0.24

edin

-0.22

idas

-0.22

yip

-0.22

tha

-0.22

nen

-0.21

oute

-0.21

POSITIVE LOGITS

wealth

0.21

 aspiring

0.20

renheit

0.20

 glamorous

0.19

 entrepreneur

0.18

 workplaces

0.18

 entrepreneurs

0.18

 fledgling

0.18

iola

0.18

 instruments

0.17

Activations Density 21.304%

numeric values

No Comments

No Known Activations