INDEX

Explanations

phrases related to potential risks or consequences

oai_token-act-pair · gpt-3.5-turbo

themes related to actions, political maneuvers, and potential consequences or risks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

ITNESS

-0.76

Joined

-0.65

 Annotations

-0.65

RIP

-0.61

andom

-0.59

DX

-0.58

Reader

-0.57

anted

-0.57

Registered

-0.56

Rated

-0.56

POSITIVE LOGITS

 would

1.52

would

1.41

 wouldn

1.36

 could

1.26

could

1.15

 undermines

1.15

 poses

1.11

 might

1.10

 entails

1.09

 will

1.08

Activations Density 0.333%

phrases related to potential risks or consequences

themes related to actions, political maneuvers, and potential consequences or risks

No Comments

No Known Activations