INDEX

Explanations

links or references to additional information

oai_token-act-pair · gpt-3.5-turbo

mentions of legal or investigative actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ãĤ¼ãĤ¦ãĤ¹

-0.78

 masc

-0.72

 mechanically

-0.70

 charms

-0.70

 scrim

-0.68

 mascara

-0.67

utic

-0.65

atu

-0.64

 attraction

-0.63

 trophies

-0.62

POSITIVE LOGITS

 Politico

1.41

 POLITICO

1.32

NYT

1.29

 Breitbart

1.25

Bloomberg

1.22

 Huffington

1.22

Polit

1.18

 Krugman

1.15

 Salon

1.10

 Brookings

1.10

Activations Density 0.581%

links or references to additional information

mentions of legal or investigative actions

No Comments

No Known Activations