INDEX

Explanations

words related to defining or describing something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "define" or its variations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 torrent

-0.64

 Trader

-0.63

 fumes

-0.62

 Alleg

-0.61

roy

-0.61

Tos

-0.60

 scam

-0.60

 rushes

-0.59

HK

-0.59

armed

-0.59

POSITIVE LOGITS

 define

3.46

 defines

2.49

 defined

2.16

define

2.13

 defining

2.13

defined

1.86

 redef

1.74

 characterize

1.69

 definition

1.69

 definitions

1.68

Activations Density 0.023%

words related to defining or describing something

instances of the word "define" or its variations

No Comments

No Known Activations