INDEX

Explanations

instances of the word "a"

oai_token-act-pair · gpt-3.5-turbo

instances of the article "a" or similar indefinite articles

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

 withd

-0.59

"}],"

-0.59

itor

-0.58

DragonMagazine

-0.58

ulous

-0.57

atis

-0.56

jit

-0.56

 "$:/

-0.55

irin

-0.53

esan

-0.53

POSITIVE LOGITS

1.49

an

1.28

 another

1.15

 something

0.97

 some

0.85

the

0.83

 someone

0.77

any

0.72

 alot

0.71

 their

0.70

Activations Density 1.030%

instances of the word "a"

instances of the article "a" or similar indefinite articles

No Comments

No Known Activations