INDEX

Explanations

financial or economic terms, specifically focused on the concept of "junk" status or junk-rated investments

oai_token-act-pair · gpt-3.5-turbo

terms related to low-quality or worthless items

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

 unintended

-0.68

APH

-0.67

SPONSORED

-0.66

 Sinai

-0.62

 Verge

-0.60

 brim

-0.58

ribe

-0.57

uve

-0.57

 Jinn

-0.57

voy

-0.56

POSITIVE LOGITS

irk

1.12

ernel

1.06

buster

1.06

ies

0.99

alos

0.95

regate

0.92

ie

0.92

tion

0.91

geon

0.91

adel

0.89

Activations Density 0.052%

financial or economic terms, specifically focused on the concept of "junk" status or junk-rated investments

terms related to low-quality or worthless items

No Comments

No Known Activations