INDEX

Explanations

phrases that indicate a change in speaker or subject in a dialogue or text

oai_token-act-pair · gpt-3.5-turbo

various forms of punctuation or non-standard characters in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 lodging

-0.66

 bunk

-0.66

 peripher

-0.64

 housed

-0.64

 peril

-0.63

 CLSID

-0.62

 scatter

-0.61

 fashioned

-0.61

 diffusion

-0.60

 dispers

-0.59

POSITIVE LOGITS

º

1.37

¹

1.23

£

1.23

ª

1.20

»

1.14

Ĵ

1.14

Į

1.14

į

1.12

¡

1.12

§

1.10

Activations Density 0.291%

phrases that indicate a change in speaker or subject in a dialogue or text

various forms of punctuation or non-standard characters in text

No Comments

No Known Activations