INDEX

Explanations

phrases and acronyms related to business or deals

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

å§«

-0.77

 Unch

-0.72

ä»£

-0.67

è£ħ

-0.63

 Amend

-0.63

éĹĺ

-0.61

 Webster

-0.61

 Dempsey

-0.59

Ĥª

-0.58

 Barcl

-0.57

POSITIVE LOGITS

seless

0.91

thing

0.90

ciples

0.88

chers

0.83

ring

0.80

assic

0.80

pling

0.78

itably

0.78

ches

0.77

acial

0.76

Activations Density 0.043%

phrases and acronyms related to business or deals

No Comments

No Known Activations