INDEX

Explanations

time-related words related to a specific time duration

oai_token-act-pair · gpt-3.5-turbo

phrases indicating quantity or number, especially relating to two or more entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

antry

-0.76

 Expression

-0.68

BIL

-0.66

indu

-0.65

 Berger

-0.65

ahime

-0.65

 Awareness

-0.63

BDS

-0.63

entimes

-0.63

 FANTASY

-0.62

POSITIVE LOGITS

 apiece

1.08

 totaling

0.90

ĪĴ

0.81

 brothers

0.77

handled

0.76

milo

0.75

undred

0.72

brother

0.72

 sisters

0.72

iliated

0.71

Activations Density 0.509%

time-related words related to a specific time duration

phrases indicating quantity or number, especially relating to two or more entities

No Comments

No Known Activations