INDEX

Explanations

words related to the brain 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_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

 Hitman

-0.69

FT

-0.69

lihood

-0.69

Else

-0.63

 Kerry

-0.62

 artisan

-0.60

anded

-0.58

makers

-0.58

 Lennon

-0.58

itsch

-0.57

POSITIVE LOGITS

ricular

1.51

urous

1.25

ilated

1.24

uring

1.20

ral

1.19

rue

1.18

ilation

1.12

uri

1.02

orial

1.00

ories

1.00

Activations Density 0.019%

words related to the brain and medical conditions

No Comments

No Known Activations