INDEX

Explanations

words related to the Pentagon

oai_token-act-pair · gpt-3.5-turbo

mentions of the Pentagon and related entities

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

âĿ

-0.81

 respecting

-0.70

fine

-0.66

Ã¡s

-0.66

sync

-0.63

 colours

-0.61

 cared

-0.60

 watering

-0.60

 colour

-0.59

vous

-0.59

POSITIVE LOGITS

 Pentagon

3.65

DOD

1.80

 Lockheed

1.51

 Defense

1.49

 USAF

1.47

 Kremlin

1.46

 Mattis

1.40

Military

1.37

military

1.28

Defense

1.27

Activations Density 0.012%

words related to the Pentagon

mentions of the Pentagon and related entities

No Comments

No Known Activations