INDEX

Explanations

mentions of royalty or royal-related terms

oai_token-act-pair · gpt-3.5-turbo

references to royal figures or royalty

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Extend

-0.83

geist

-0.82

LER

-0.80

 Loaded

-0.77

ulhu

-0.77

onica

-0.77

 Glock

-0.75

 Apply

-0.75

 Canaver

-0.74

Spread

-0.73

POSITIVE LOGITS

 decree

1.08

 pard

1.02

 palace

0.95

 royal

0.94

 princes

0.89

 immunity

0.89

 decre

0.87

 prince

0.87

pin

0.84

 royalty

0.83

Activations Density 0.011%

mentions of royalty or royal-related terms

references to royal figures or royalty

No Comments

No Known Activations