INDEX

Explanations

words related to cyber attacks and military actions

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

-0.87

soDeliveryDate

-0.83

aird

-0.79

BuyableInstoreAndOnline

-0.78

chart

-0.71

paren

-0.71

venture

-0.71

gio

-0.68

hall

-0.67

uid

-0.67

POSITIVE LOGITS

 innocent

1.16

 unsuspecting

1.16

 havoc

1.00

 civilians

1.00

 innoc

0.89

 unprotected

0.88

 targets

0.87

 indiscrim

0.82

him

0.81

 unarmed

0.81

Activations Density 4.737%

words related to cyber attacks and military actions

No Comments

No Known Activations