INDEX

Explanations

phrases related to official orders or proclamations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ortmund

-0.81

doms

-0.75

regate

-0.74

idences

-0.72

 Flavoring

-0.71

çīĪ

-0.70

 Increases

-0.70

arms

-0.69

 VIDEOS

-0.68

ipeg

-0.67

POSITIVE LOGITS

 supposed

1.00

 considered

0.91

 going

0.89

 aware

0.89

 gonna

0.87

 ready

0.86

 barred

0.85

 concerned

0.83

 ashamed

0.82

 able

0.82

Activations Density 2.014%

phrases related to official orders or proclamations

No Comments

No Known Activations