INDEX

Explanations

phrases related to the concept of "means" or methods of doing something

oai_token-act-pair · gpt-3.5-turbo

phrases related to methods or means of various activities or processes

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

bats

-0.83

fml

-0.79

warm

-0.78

nis

-0.77

lights

-0.76

stones

-0.76

busters

-0.75

kids

-0.74

 Freak

-0.72

 Shades

-0.71

POSITIVE LOGITS

 obtaining

1.14

 accessing

1.07

 circumvent

1.06

 transportation

1.06

 extracting

1.03

 determining

1.03

 communication

1.02

 achieving

0.99

 communicate

0.98

 distributing

0.98

Activations Density 0.209%

phrases related to the concept of "means" or methods of doing something

phrases related to methods or means of various activities or processes

No Comments

No Known Activations