INDEX

Explanations

verbs related to daily activities and routines, as well as descriptive words that indicate emotions or actions

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

BuyableInstoreAndOnline

-0.74

Winged

-0.67

æľ

-0.66

Lenin

-0.61

ĸļ

-0.61

é¾įå¥ĳå£«

-0.60

idelines

-0.60

Balt

-0.59

Baltimore

-0.58

ª

-0.58

POSITIVE LOGITS

 safely

1.14

 easily

1.13

 anytime

0.95

 comfortably

0.95

 freely

0.94

 cheaply

0.93

 anywhere

0.91

 indefinitely

0.90

 anonymously

0.85

 efficiently

0.82

Activations Density 2.216%

verbs related to daily activities and routines, as well as descriptive words that indicate emotions or actions

No Comments

No Known Activations