INDEX

Explanations

references to religious figures, specifically priests

oai_token-act-pair · gpt-3.5-turbo

references to priests and their roles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ø§ÙĦ

-0.76

chell

-0.74

upon

-0.71

DonaldTrump

-0.71

Newsletter

-0.70

ilater

-0.70

 Flavoring

-0.69

BAT

-0.68

PsyNetMessage

-0.68

-0.67

POSITIVE LOGITS

esses

1.47

 priests

1.28

 priest

1.27

ess

1.04

 exorc

0.94

 Priest

0.90

osate

0.88

 ordained

0.87

 clergy

0.85

 priesthood

0.84

Activations Density 0.007%

references to religious figures, specifically priests

references to priests and their roles

No Comments

No Known Activations