INDEX

Explanations

terms related to various political and geographical entities, potentially related to conflict and border regions

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

anwhile

-0.43

 conclud

-0.41

 destro

-0.40

 disadvant

-0.38

 behav

-0.38

 Vaugh

-0.36

 Azerb

-0.36

 undermin

-0.36

 challeng

-0.36

 explan

-0.36

POSITIVE LOGITS

 âĢº

0.37

 Belfast

0.34

erville

0.31

Äĵ

0.28

brew

0.28

own

0.28

\":

0.27

cipled

0.27

ickr

0.27

ode

0.26

Activations Density 10.648%

terms related to various political and geographical entities, potentially related to conflict and border regions

No Comments

No Known Activations