INDEX

Explanations

phrases related to starting a fire

oai_token-act-pair · gpt-3.5-turbo

references to ignition and combustion processes in various contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

teness

-0.83

å¥³

-0.80

merce

-0.79

worthiness

-0.79

shape

-0.76

manship

-0.73

kb

-0.73

ibilities

-0.72

onomy

-0.72

omics

-0.71

POSITIVE LOGITS

 stove

1.68

 furnace

1.57

 oven

1.56

 fireplace

1.48

 heater

1.48

 torches

1.41

 cooker

1.36

 candles

1.33

 furn

1.31

 lamps

1.31

Activations Density 0.361%

phrases related to starting a fire

references to ignition and combustion processes in various contexts

No Comments

No Known Activations