INDEX

Explanations

acronyms "AC" with digit values associated with them

oai_token-act-pair · gpt-3.5-turbo

references to AC (alternating current) and possibly other electrical terms

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

glers

-0.72

 Tradable

-0.69

hs

-0.67

hung

-0.67

hood

-0.64

chens

-0.64

jar

-0.63

lihood

-0.62

ozyg

-0.61

 backstage

-0.61

POSITIVE LOGITS

ORN

1.35

CEPT

1.18

LU

1.09

UTE

1.02

HE

1.01

KN

1.01

OSS

0.97

APTER

0.97

ARS

0.95

ULAR

0.95

Activations Density 0.020%

acronyms "AC" with digit values associated with them

references to AC (alternating current) and possibly other electrical terms

No Comments

No Known Activations