INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

references to a specific individual or entity with the name "Arne."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ĸļ

-1.04

¬¼

-0.79

isSpecialOrderable

-0.72

hower

-0.70

eners

-0.68

stakes

-0.67

ĨĴ

-0.65

 smokes

-0.65

 sterling

-0.62

éĹĺ

-0.62

POSITIVE LOGITS

beit

1.13

ansas

1.12

rival

1.09

ithmetic

1.01

paio

0.99

duino

0.98

lington

0.95

rived

0.95

rington

0.95

issa

0.93

Activations Density 0.018%

names of individuals

references to a specific individual or entity with the name "Arne."

No Comments

No Known Activations