INDEX

Explanations

instances of the word "automatically."

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

IFrame

Link

Negative Logits

ĸļ

-0.93

ippi

-0.80

Liter

-0.76

andals

-0.73

rug

-0.71

ergus

-0.70

 Mouth

-0.70

 Emin

-0.68

,,,,

-0.66

ilion

-0.64

POSITIVE LOGITS

 grav

0.89

 populate

0.87

 migrate

0.86

 regenerate

0.86

 detect

0.85

upd

0.84

 induct

0.82

 deducted

0.82

 detects

0.80

 combust

0.79

Activations Density 0.028%

instances of the word "automatically."

No Known Activations