INDEX

Explanations

the word 'le'

oai_token-act-pair · gpt-3.5-turbo

occurrences of the prefix "le-"

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

akeru

-0.72

DCS

-0.70

Wan

-0.69

 Imam

-0.69

Discover

-0.65

USD

-0.65

ogun

-0.64

 Reson

-0.63

obook

-0.62

iment

-0.61

POSITIVE LOGITS

le

3.62

Le

1.74

Le

1.56

les

1.33

 Leopard

1.27

LE

1.25

le

1.20

 lemon

1.18

la

1.17

Lem

1.16

Activations Density 0.017%

the word 'le'

occurrences of the prefix "le-"

No Comments

No Known Activations