INDEX

Explanations

names of websites and online platforms

oai_token-act-pair · gpt-3.5-turbo

phrases indicating locations or associations with specific organizations and individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ETF

-0.77

ãĤ¼ãĤ¦ãĤ¹

-0.70

 dressing

-0.69

 mechanically

-0.67

 hinges

-0.64

utic

-0.63

isol

-0.62

artifacts

-0.61

 compounds

-0.61

Catalog

-0.61

POSITIVE LOGITS

 Slate

1.19

 Huffington

1.18

 Wired

1.14

 Forbes

1.14

 Politico

1.13

 Gawker

1.13

Vox

1.07

 Bloomberg

1.03

blogs

1.02

 Breitbart

1.02

Activations Density 0.283%

names of websites and online platforms

phrases indicating locations or associations with specific organizations and individuals

No Comments

No Known Activations