INDEX

Explanations

the name "Fahmy"

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals and names

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

sed

-0.71

 port

-0.71

 bananas

-0.71

rip

-0.71

DOS

-0.68

xit

-0.67

BAS

-0.67

 ports

-0.65

 Tropical

-0.65

 Roland

-0.64

POSITIVE LOGITS

 Leah

1.99

Fe

1.88

Fah

1.68

Sed

1.31

 Feldman

1.24

ICH

1.16

 Scha

1.12

Wee

1.01

 Feinstein

0.92

CLA

0.91

Activations Density 0.054%

the name "Fahmy"

references to specific individuals and names

No Comments

No Known Activations