INDEX

Explanations

phrases related to receipts or receiving items

oai_token-act-pair · gpt-3.5-turbo

mentions of receipts and related documentation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Hosp

-0.66

adders

-0.62

 Penal

-0.60

 Capital

-0.59

croft

-0.59

unin

-0.58

RH

-0.57

projects

-0.56

ovych

-0.56

roup

-0.55

POSITIVE LOGITS

 receipt

1.27

 receipts

1.23

prise

0.76

ãĥ¼ãĥĨãĤ£

0.75

oire

0.74

inosaur

0.72

 REPORT

0.72

TY

0.71

bable

0.70

orial

0.70

Activations Density 0.008%

phrases related to receipts or receiving items

mentions of receipts and related documentation

No Comments

No Known Activations