INDEX

Explanations

terms related to military aircraft and air traffic control

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.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

 tuition

-0.85

ucket

-0.80

galitarian

-0.76

 millionaires

-0.75

hemy

-0.74

stros

-0.74

itutional

-0.74

lesh

-0.74

rylic

-0.72

 vegan

-0.72

POSITIVE LOGITS

 radar

1.48

 signals

1.45

 transmissions

1.44

 alerts

1.44

 alert

1.40

 signal

1.39

 Alert

1.36

 transmitting

1.34

 radios

1.30

alert

1.28

Activations Density 0.559%

terms related to military aircraft and air traffic control

No Comments

No Known Activations