INDEX

Explanations

expressions of personal reaction and ownership

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

ournals

-0.71

idelines

-0.68

 Printed

-0.66

soDeliveryDate

-0.65

pub

-0.65

etary

-0.61

 onto

-0.61

Dialogue

-0.61

artifacts

-0.60

chairs

-0.59

POSITIVE LOGITS

 dismay

1.27

 surprise

1.19

 extent

1.14

 aston

1.13

 amaz

1.09

 detriment

1.07

 annoyance

1.02

 delight

1.01

 disappointment

0.95

 puzz

0.94

Activations Density 0.052%

expressions of personal reaction and ownership

No Comments

No Known Activations