INDEX

Explanations

numerical values related to property features

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

ragon

-0.80

ozyg

-0.77

aclysm

-0.75

Transaction

-0.74

ocious

-0.72

otide

-0.72

 milliseconds

-0.70

ERROR

-0.70

 Reck

-0.69

DNA

-0.68

POSITIVE LOGITS

 seating

1.42

 bedrooms

1.27

 rooms

1.25

 entrances

1.24

 spacious

1.23

 balcon

1.22

 ventilation

1.22

 balcony

1.21

 lounge

1.20

 adjoining

1.20

Activations Density 0.591%

numerical values related to property features

No Comments

No Known Activations