INDEX

Explanations

references to the release or launch of products or services, particularly in the technology sector

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

acca

-0.73

PI

-0.70

ukong

-0.68

rill

-0.67

oras

-0.66

iery

-0.66

insula

-0.65

 Travels

-0.65

igate

-0.64

ornia

-0.64

POSITIVE LOGITS

 nonetheless

1.83

 nevertheless

1.59

etheless

1.31

 still

1.02

 alas

0.99

 cautioned

0.94

 lacked

0.92

 beware

0.91

 lacks

0.89

 retains

0.89

Activations Density 2.679%

references to the release or launch of products or services, particularly in the technology sector

No Comments

No Known Activations