INDEX

Explanations

references to the concept of lighting a candle

oai_token-act-pair · gpt-3.5-turbo

references to candles and light sources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

doms

-0.84

ese

-0.82

omal

-0.73

ele

-0.71

FactoryReloaded

-0.69

laus

-0.68

ocard

-0.68

erion

-0.68

ettings

-0.67

ordan

-0.66

POSITIVE LOGITS

 candle

1.07

lit

1.06

 candles

1.01

light

0.99

 Candle

0.97

 illuminating

0.93

 lantern

0.92

lit

0.84

 illuminated

0.81

 illuminate

0.81

Activations Density 0.031%

references to the concept of lighting a candle

references to candles and light sources

No Comments

No Known Activations