INDEX

Explanations

personal pronouns followed by a statement

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

iquette

-0.62

 toget

-0.61

ãĤ¼

-0.60

 Hels

-0.60

Redditor

-0.59

eatures

-0.59

redients

-0.57

 è£ıè

-0.56

 Abyss

-0.55

pires

-0.54

POSITIVE LOGITS

 think

1.42

'm

1.37

 mean

1.25

've

1.18

 guess

1.17

don

1.16

 suppose

1.05

 dunno

1.04

'd

1.02

 wouldn

1.02

Activations Density 0.177%

personal pronouns followed by a statement

No Comments

No Known Activations