INDEX

Explanations

pairs of words used in technical contexts, such as "voltage pins" and "signal wires."

oai_token-act-pair · gpt-3.5-turbo

conjunctions and phrases relating to connections and relationships between elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

political

-0.75

justice

-0.75

acus

-0.75

OWN

-0.73

minent

-0.72

massive

-0.72

headed

-0.71

landish

-0.71

meet

-0.70

ander

-0.69

POSITIVE LOGITS

 optionally

1.59

 thus

1.32

 hence

1.31

 then

1.28

 thereby

1.24

 therefore

1.23

 THEN

1.21

 consequently

1.21

 thence

1.17

 vice

1.16

Activations Density 0.261%

pairs of words used in technical contexts, such as "voltage pins" and "signal wires."

conjunctions and phrases relating to connections and relationships between elements

No Comments

No Known Activations