INDEX

Explanations

phrases related to political news and government policies

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

audi

-0.70

amiya

-0.63

overe

-0.62

ilial

-0.58

 Neurolog

-0.56

 behalf

-0.55

CRE

-0.55

 deficits

-0.55

aring

-0.54

oster

-0.54

POSITIVE LOGITS

 traction

0.93

 foothold

0.90

 mileage

0.77

 bearings

0.77

chy

0.73

 FREE

0.70

 footing

0.69

ãĥĸ

0.69

 attention

0.69

started

0.67

Activations Density 2.201%

phrases related to political news and government policies

No Comments

No Known Activations