INDEX

Explanations

phrases indicating a lack of connection or value

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

hooting

-0.81

FIG

-0.66

yond

-0.65

merce

-0.64

itsch

-0.63

duction

-0.63

inet

-0.62

eper

-0.62

 Posts

-0.62

Sport

-0.61

POSITIVE LOGITS

 knack

1.15

 chance

1.06

 opportunity

0.93

 vested

0.93

 tendency

0.92

been

0.86

 penchant

0.86

 clout

0.85

 affinity

0.84

 reputation

0.83

Activations Density 2.175%

phrases indicating a lack of connection or value

No Comments

No Known Activations