INDEX

Explanations

mentions of or related to supreme leadership or authority figures

oai_token-act-pair · gpt-3.5-turbo

references to supreme leaders and authorities

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

phia

-0.99

strip

-0.75

lishes

-0.74

Äĩ

-0.72

glass

-0.69

dream

-0.67

vier

-0.66

thening

-0.64

oney

-0.63

itten

-0.63

POSITIVE LOGITS

 Court

1.24

Ct

1.01

 Tribunal

1.00

 court

0.97

Court

0.93

 Commander

0.88

 Allied

0.87

 COURT

0.86

 Judicial

0.85

 Ruler

0.82

Activations Density 0.028%

mentions of or related to supreme leadership or authority figures

references to supreme leaders and authorities

No Comments

No Known Activations