INDEX

Explanations

words related to creating, updating, or modifying things

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

OPA

-0.77

acebook

-0.75

xual

-0.73

AFTA

-0.70

phe

-0.67

 Coulter

-0.67

ONSORED

-0.67

 Nieto

-0.66

 Limbaugh

-0.66

SPONSORED

-0.66

POSITIVE LOGITS

bie

1.46

bies

1.33

foundland

0.98

 batch

0.95

 arrivals

0.88

egg

0.88

 Zealand

0.85

castle

0.80

 generation

0.78

clamation

0.76

Activations Density 0.069%

words related to creating, updating, or modifying things

No Comments

No Known Activations