INDEX

Explanations

characters' names, titles, and place names from a specific movie or series

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

 constitu

-0.62

terday

-0.62

 confidence

-0.61

 Bris

-0.59

 Hera

-0.57

 WATCHED

-0.56

 Juliet

-0.56

OIL

-0.56

 Remem

-0.55

Io

-0.55

POSITIVE LOGITS

etermin

1.29

ynam

1.26

ividual

1.24

iamond

1.23

ynamic

1.22

imensional

1.19

iverse

1.18

etermination

1.17

etermined

1.16

iscovery

1.14

Activations Density 3.615%

characters' names, titles, and place names from a specific movie or series

No Comments

No Known Activations