INDEX

Explanations

questions or statements about the identity or actions of individuals

oai_token-act-pair · gpt-3.5-turbo

instances of the phrase "who's" as well as related questions about identity and responsibility

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

avascript

-0.79

nothing

-0.71

tesque

-0.70

earchers

-0.67

olson

-0.65

APD

-0.65

urances

-0.63

 Heights

-0.63

iosity

-0.62

obi

-0.62

POSITIVE LOGITS

 responsible

1.01

 footing

0.97

 whom

0.95

 hardest

0.88

 closest

0.87

 benefiting

0.86

 accountable

0.84

responsible

0.83

 fault

0.81

 sponsoring

0.81

Activations Density 0.156%

questions or statements about the identity or actions of individuals

instances of the phrase "who's" as well as related questions about identity and responsibility

No Comments

No Known Activations