INDEX

Explanations

medical terms and concepts related to operations, surgical techniques, and medical conditions

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

ochond

-0.73

arlane

-0.71

ortment

-0.70

usalem

-0.69

ĪĴ

-0.65

weeney

-0.63

odan

-0.61

cipl

-0.60

apons

-0.60

icrobial

-0.60

POSITIVE LOGITS

 unless

0.92

 anyway

0.91

0.86

 because

0.84

 anyways

0.82

 anymore

0.81

0.78

.—

0.77

.",

0.77

 nowadays

0.76

Activations Density 1.771%

medical terms and concepts related to operations, surgical techniques, and medical conditions

No Comments

No Known Activations