INDEX

Explanations

quantitative values or measurements

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

iolet

-0.61

arna

-0.56

onal

-0.56

rib

-0.55

 conclud

-0.55

arcity

-0.53

heed

-0.52

terday

-0.52

akia

-0.52

 Release

-0.51

POSITIVE LOGITS

liest

0.91

of

0.89

same

0.86

 thereof

0.79

iest

0.79

forts

0.77

achu

0.73

hest

0.70

antry

0.67

est

0.67

Activations Density 1.378%

quantitative values or measurements

No Comments

No Known Activations