INDEX

Explanations

references to actions and abilities

oai_token-act-pair · gpt-3.5-turbo

phrases related to actions and capabilities of subjects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

QB

-0.68

Border

-0.68

Road

-0.66

 Avalon

-0.66

Nation

-0.63

 Nirvana

-0.63

 Cabin

-0.62

é¾

-0.62

 Aval

-0.62

Track

-0.60

POSITIVE LOGITS

 nonetheless

1.48

 nevertheless

1.45

 certainly

0.92

 anecd

0.82

 occasionally

0.81

etheless

0.77

 indirectly

0.74

 anyway

0.72

 suffice

0.70

 inferred

0.70

Activations Density 0.475%

references to actions and abilities

phrases related to actions and capabilities of subjects

No Comments

No Known Activations