INDEX

Explanations

phrases or sentences that involve the concept of two related things or ideas

oai_token-act-pair · gpt-3.5-turbo

phrases indicating a connection or relationship between two elements

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

any

-0.71

eva

-0.68

 name

-0.66

ews

-0.65

old

-0.62

urg

-0.60

agate

-0.60

ots

-0.59

atz

-0.59

iz

-0.58

POSITIVE LOGITS

 coupled

3.76

 paired

2.17

 combined

1.85

 contrasted

1.47

 coupling

1.47

 compounded

1.39

 supplemented

1.31

 sandwic

1.25

 accompanied

1.23

 correlated

1.16

Activations Density 0.024%

phrases or sentences that involve the concept of two related things or ideas

phrases indicating a connection or relationship between two elements

No Comments

No Known Activations