INDEX

Explanations

phrases related to cybercrime and allegations of hacking

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĤ´ãĥ³

-0.86

utenberg

-0.77

HUD

-0.76

baugh

-0.74

rea

-0.72

axis

-0.70

rite

-0.69

ãĥ¯ãĥ³

-0.68

zar

-0.67

rition

-0.67

POSITIVE LOGITS

 unsuspecting

1.25

 innocent

0.89

 unwitting

0.88

 havoc

0.82

 civilians

0.80

 humans

0.79

 Australians

0.77

 hostages

0.75

 computers

0.74

us

0.73

Activations Density 0.301%

phrases related to cybercrime and allegations of hacking

No Comments

No Known Activations