INDEX

Explanations

details related to the word "Nat"

oai_token-act-pair · gpt-3.5-turbo

references to the term "Nat" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

wordpress

-0.68

 Bearing

-0.67

dress

-0.65

inventoryQuantity

-0.64

 Peninsula

-0.64

STON

-0.64

 contraceptives

-0.64

ĳå£«

-0.63

poon

-0.62

deen

-0.62

POSITIVE LOGITS

asha

1.02

ured

1.01

ivas

1.00

ritional

0.99

omas

0.96

ual

0.94

itionally

0.91

uro

0.88

Geo

0.88

olith

0.88

Activations Density 0.033%

details related to the word "Nat"

references to the term "Nat" in various contexts

No Comments

No Known Activations