INDEX

Explanations

phrases related to royalty or prominent figures

oai_token-act-pair · gpt-3.5-turbo

phrases indicating authority or prominence, particularly related to titles or roles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

annis

-0.68

llah

-0.67

matic

-0.64

llor

-0.62

rouse

-0.62

eful

-0.61

urches

-0.61

Pope

-0.61

 directly

-0.60

bishop

-0.60

POSITIVE LOGITS

 sorts

0.86

 mankind

0.82

 Gamergate

0.73

 Mankind

0.72

 nerds

0.70

 humankind

0.70

 hearts

0.70

 modern

0.68

 understatement

0.67

rap

0.66

Activations Density 0.156%

phrases related to royalty or prominent figures

phrases indicating authority or prominence, particularly related to titles or roles

No Comments

No Known Activations