INDEX

Explanations

instructions or steps on how to make something

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

uploads

-0.67

eatures

-0.67

lights

-0.66

blogspot

-0.62

court

-0.62

bryce

-0.61

 entirety

-0.61

axter

-0.59

ipped

-0.58

ippy

-0.57

POSITIVE LOGITS

 efficiently

0.80

uate

0.77

 yourself

0.76

 oneself

0.74

 safely

0.70

 Yourself

0.65

 your

0.65

 interact

0.65

ocate

0.65

 attribution

0.64

Activations Density 0.266%

instructions or steps on how to make something

No Comments

No Known Activations