INDEX

Explanations

words related to various types of products, activities, and events

oai_token-act-pair · gpt-3.5-turbo

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

--------------------------------------------------------

-0.77

Tokens

-0.67

Ĭ±

-0.63

actionDate

-0.62

SSL

-0.60

 Roof

-0.60

CLR

-0.59

KNOWN

-0.58

 Users

-0.56

cknow

-0.55

POSITIVE LOGITS

 represents

1.24

 belongs

1.21

 deserves

1.18

 reminds

1.12

 signifies

1.10

 proves

1.09

 serves

1.09

 contains

1.09

 differs

1.07

 demonstrates

1.06

Activations Density 0.157%

words related to various types of products, activities, and events

No Comments

No Known Activations