INDEX

Explanations

uncommon acronyms and abbreviations

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

nces

-0.65

 Mechanics

-0.65

 Athletic

-0.63

OPLE

-0.62

 Madness

-0.61

 Passenger

-0.61

senal

-0.58

 ashore

-0.57

Accessory

-0.56

)=(

-0.56

POSITIVE LOGITS

ield

1.33

riend

1.32

ornia

1.29

amily

1.21

rost

1.21

unction

1.18

eatures

1.16

actory

1.14

lex

1.14

onso

1.13

Activations Density 0.644%

uncommon acronyms and abbreviations

No Comments

No Known Activations