INDEX

Explanations

numbers related to statistical data and figures

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

 parach

-0.64

 aside

-0.63

 advis

-0.62

 traveler

-0.62

 predatory

-0.62

 proactive

-0.62

 shroud

-0.62

 pregnancy

-0.62

 swim

-0.61

 proposition

-0.61

POSITIVE LOGITS

1.47

1.41

1.33

1.31

1.30

1.29

1.28

1.25

Activations Density 0.543%

numbers related to statistical data and figures

No Comments

No Known Activations