INDEX

Explanations

mentions of various locations

oai_token-act-pair · gpt-3.5-turbo

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

 mileage

-0.68

Downloadha

-0.66

 Galile

-0.66

 slic

-0.65

 stitches

-0.61

Reviewer

-0.61

saf

-0.59

ĪĴ

-0.58

NIC

-0.58

 laundry

-0.58

POSITIVE LOGITS

hirt

1.37

weet

1.21

arnaev

1.19

leeve

1.17

onic

1.05

hower

1.04

erker

1.00

cape

1.00

ugar

0.97

pec

0.96

Activations Density 0.054%

mentions of various locations

No Comments

No Known Activations