INDEX

Explanations

mentions of political figures and countries

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

 Bronx

-0.83

 Bucc

-0.82

emetery

-0.81

 Darrell

-0.80

 Died

-0.78

 anecd

-0.74

ped

-0.73

 Loft

-0.71

Adult

-0.71

 SCHOOL

-0.71

POSITIVE LOGITS

 Nato

1.46

 NATO

1.45

 diplomatic

1.40

 treaties

1.31

 diplomats

1.30

 treaty

1.29

 diplomacy

1.27

 bilateral

1.27

 Treaty

1.26

 milit

1.25

Activations Density 3.589%

mentions of political figures and countries

No Comments

No Known Activations