INDEX

Explanations

political and governmental terms and actions

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

ãĤ¼

-0.75

photos

-0.72

 Splash

-0.72

Bott

-0.70

Afee

-0.69

Cry

-0.69

Atl

-0.68

Jet

-0.67

itars

-0.67

Pure

-0.67

POSITIVE LOGITS

 overseeing

1.45

 oversee

1.40

 convened

1.40

 jurisdiction

1.35

 oversight

1.33

 appoint

1.31

 appointed

1.30

 oversees

1.29

 chaired

1.26

 presiding

1.25

Activations Density 3.503%

political and governmental terms and actions

No Comments

No Known Activations