INDEX

Explanations

locations or geographic regions

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

 Vaugh

-0.55

 Niet

-0.46

enegger

-0.45

 undermin

-0.44

 Moroc

-0.43

 Instr

-0.42

 notor

-0.41

 Azerb

-0.40

 advoc

-0.40

retty

-0.40

POSITIVE LOGITS

Â¶

0.57

 âĢº

0.54

):

0.53

)?

0.52

 Belfast

0.47

↵

0.47

 Posted

0.44

↵↵

0.42

 âĢ¢

0.41

Activations Density 16.809%

locations or geographic regions

No Comments

No Known Activations