INDEX

Explanations

medical terms and symptoms

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

edia

-0.74

ahime

-0.66

odcast

-0.64

senal

-0.64

resy

-0.63

tch

-0.62

eele

-0.62

uristic

-0.61

maxwell

-0.60

eka

-0.60

POSITIVE LOGITS

 except

1.70

except

1.21

 imaginable

1.13

 alike

1.13

 whatsoever

1.11

 irrespective

1.10

soever

1.05

 regardless

0.99

 thereafter

0.92

 besides

0.86

Activations Density 0.244%

medical terms and symptoms

No Comments

No Known Activations