INDEX

Explanations

questions starting with "What" or "How"

oai_token-act-pair · gpt-3.5-turbo

questions or prompts that invite further exploration or inquiry

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

rique

-0.64

unction

-0.64

udos

-0.63

 Settlement

-0.63

cott

-0.62

¬¼

-0.62

minster

-0.62

wash

-0.61

oppers

-0.61

§

-0.60

POSITIVE LOGITS

 namely

1.14

viz

0.92

 whether

0.89

 versus

0.87

etc

0.86

whether

0.85

 realistically

0.84

how

0.84

 besides

0.84

 assuming

0.81

Activations Density 0.279%

questions starting with "What" or "How"

questions or prompts that invite further exploration or inquiry

No Comments

No Known Activations