INDEX

Explanations

references to postage stamps

oai_token-act-pair · gpt-3.5-turbo

references to postage stamps

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

onomous

-0.71

 Heights

-0.71

annis

-0.67

accessible

-0.66

ther

-0.66

 Ther

-0.65

ors

-0.65

involved

-0.63

connected

-0.63

inis

-0.63

POSITIVE LOGITS

 stamp

4.06

 stamps

3.45

 Stamp

2.64

 stamped

2.04

 imprint

1.36

 sticker

1.35

 postage

1.31

 seal

1.17

 stomp

1.10

 stickers

1.03

Activations Density 0.012%

references to postage stamps

references to postage stamps

No Comments

No Known Activations