INDEX

Explanations

references to personal information or activities

oai_token-act-pair · gpt-3.5-turbo

references to personal matters or experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

xual

-1.03

ERY

-0.74

 Cuomo

-0.71

peak

-0.68

UG

-0.68

LOS

-0.68

 Nile

-0.67

 Kenny

-0.67

 Archdemon

-0.67

REG

-0.67

POSITIVE LOGITS

 belongings

1.05

personal

0.95

 personal

0.95

ised

0.94

 estate

0.88

 wellbeing

0.87

 pronouns

0.84

estate

0.83

isations

0.81

istically

0.80

Activations Density 0.016%

references to personal information or activities

references to personal matters or experiences

No Comments

No Known Activations