INDEX

Explanations

hashtags

oai_token-act-pair · gpt-3.5-turbo

references to numerical rankings and positions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 destro

-0.86

 Gors

-0.77

ãĤ©

-0.74

 cler

-0.73

 Nanto

-0.72

oun

-0.72

gib

-0.72

 commissions

-0.71

 concentrating

-0.70

 seiz

-0.68

POSITIVE LOGITS

DIV

1.27

###

1.06

################################

1.06

########

1.05

DonaldTrump

1.01

define

0.99

!/

0.99

Gamer

0.96

################

0.95

Ask

0.93

Activations Density 0.018%

hashtags

references to numerical rankings and positions

No Comments

No Known Activations