INDEX

Explanations

mentions of interviews and dialogues between people

oai_token-act-pair · gpt-3.5-turbo

references to interviews and published articles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 footing

-0.75

 suff

-0.73

 lumber

-0.70

 overpower

-0.67

gage

-0.67

 choke

-0.67

 habitats

-0.67

 overpowered

-0.67

 strap

-0.66

ranged

-0.66

POSITIVE LOGITS

 Pastebin

0.97

 VIDEOS

0.86

 redacted

0.77

Journal

0.72

 Medline

0.71

comment

0.71

DragonMagazine

0.71

 Blog

0.70

 Paste

0.70

reports

0.70

Activations Density 1.815%

mentions of interviews and dialogues between people

references to interviews and published articles

No Comments

No Known Activations