INDEX

Explanations

locations or places

oai_token-act-pair · gpt-3.5-turbo

terms related to the theme of "surprising" or "surprisingly."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

naires

-0.86

hetti

-0.73

heid

-0.72

Constructed

-0.72

assetsadobe

-0.66

 withholding

-0.63

itsch

-0.62

 amber

-0.61

 Eternity

-0.61

 overhead

-0.61

POSITIVE LOGITS

geon

1.36

geons

1.22

rounded

1.03

render

1.01

fer

0.95

fing

0.94

round

0.94

vey

0.87

prises

0.86

ports

0.86

Activations Density 0.008%

locations or places

terms related to the theme of "surprising" or "surprisingly."

No Comments

No Known Activations