INDEX

Explanations

words related to criminal activities and investigations

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

mble

-0.88

ngth

-0.75

 Scythe

-0.75

VK

-0.73

hirt

-0.73

bler

-0.71

alth

-0.71

pread

-0.70

forth

-0.70

NetMessage

-0.69

POSITIVE LOGITS

 Torres

1.25

Ã±a

1.24

Ã±o

1.24

 JosÃ©

1.21

ÃŃa

1.19

 Cuban

1.16

Ã±

1.16

 Latino

1.15

 Honduras

1.14

 Nieto

1.14

Activations Density 0.401%

words related to criminal activities and investigations

No Comments

No Known Activations