INDEX

Explanations

images of shirts

oai_token-act-pair · gpt-3.5-turbo

references to shirts, particularly T-shirts

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

audi

-0.77

yrinth

-0.74

ntil

-0.67

judicial

-0.66

nesota

-0.62

elsen

-0.62

ruciating

-0.61

 trib

-0.61

cffffcc

-0.60

enthal

-0.60

POSITIVE LOGITS

shirt

1.21

 shirts

1.11

 sleeve

1.10

shirts

1.10

leeve

1.06

 sleeves

1.03

 shirt

1.03

hirt

1.02

rack

0.89

pins

0.88

Activations Density 0.016%

images of shirts

references to shirts, particularly T-shirts

No Comments

No Known Activations