INDEX

Explanations

phrases related to consequences and impacts of actions

oai_token-act-pair · gpt-3.5-turbo

phrases indicating negative consequences or outcomes

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

DragonMagazine

-0.81

nect

-0.77

ĸļ

-0.75

thank

-0.74

ledged

-0.73

uben

-0.73

ernaut

-0.73

Pry

-0.70

uve

-0.70

assad

-0.69

POSITIVE LOGITS

 misunderstand

1.50

 resentment

1.34

 mistrust

1.34

 distrust

1.28

 confusion

1.27

 cynicism

1.25

 counterproductive

1.25

 unnecessary

1.24

 misunderstanding

1.23

 unintended

1.22

Activations Density 0.339%

phrases related to consequences and impacts of actions

phrases indicating negative consequences or outcomes

No Comments

No Known Activations