INDEX

Explanations

mentions or discussions regarding manuscripts

oai_token-act-pair · gpt-3.5-turbo

references to manuscripts and publication-related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

evil

-0.75

 Spears

-0.73

laws

-0.70

vic

-0.69

nie

-0.66

upon

-0.66

omer

-0.66

eping

-0.65

slaught

-0.65

dh

-0.64

POSITIVE LOGITS

 manuscript

1.46

 manuscripts

1.36

uscript

0.99

 conclud

0.91

lishing

0.89

 guiActiveUn

0.82

natureconservancy

0.79

 copy

0.76

 âī¡

0.75

Activations Density 0.006%

mentions or discussions regarding manuscripts

references to manuscripts and publication-related terms

No Comments

No Known Activations