INDEX

Explanations

sequences of actions or steps

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Originally

-0.23

 Sail

-0.21

':

-0.21

 Posted

-0.20

 Splash

-0.20

HUD

-0.20

 Rollins

-0.19

 Proud

-0.19

 Fail

-0.19

 Quan

-0.19

POSITIVE LOGITS

 tradem

0.24

entimes

0.22

ÃĥÃĤ

0.22

 dilig

0.22

).[

0.21

 notor

0.21

 conflic

0.21

 challeng

0.21

amount

0.21

VIDIA

0.21

Activations Density 17.445%

sequences of actions or steps

No Comments

No Known Activations