INDEX

Explanations

mentions of specific textual locations — starting references or origins

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

merce

-0.89

faced

-0.83

ratulations

-0.83

rils

-0.75

irm

-0.74

hai

-0.73

busters

-0.73

tailed

-0.70

seek

-0.68

heet

-0.68

POSITIVE LOGITS

 afar

1.72

 whence

1.28

 scratch

1.18

 abroad

1.16

 inside

1.01

 thence

0.95

 within

0.94

 somewhere

0.92

 anywhere

0.90

 outside

0.88

Activations Density 0.546%

mentions of specific textual locations — starting references or origins

No Comments

No Known Activations