INDEX

Explanations

phrases related to politics and government

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to power dynamics and control in governance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

anooga

-0.79

GGGGGGGG

-0.74

pointers

-0.72

Missing

-0.68

Letter

-0.67

 Missing

-0.66

fresh

-0.65

emetery

-0.65

REF

-0.65

imilar

-0.65

POSITIVE LOGITS

 control

1.61

 govern

1.40

 dominate

1.38

 controlled

1.35

 controls

1.34

 wield

1.34

own

1.26

 dictate

1.22

 democratically

1.19

 CONTROL

1.17

Activations Density 0.423%

phrases related to politics and government

references to power dynamics and control in governance

No Comments

No Known Activations