INDEX

Explanations

words related to the acronym 'UD'

oai_token-act-pair · gpt-3.5-turbo

references to "UD" and related terms that denote a specific entity or concept

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

tein

-0.82

strength

-0.70

ï¸

-0.66

 Remem

-0.63

Constructed

-0.62

 Candle

-0.61

collar

-0.60

 temper

-0.60

 plates

-0.59

 Slaughter

-0.59

POSITIVE LOGITS

UD

1.28

uds

1.01

IFIED

1.00

PDATE

0.99

PLIC

0.95

iamond

0.91

BLE

0.85

anium

0.83

aeda

0.83

nesday

0.82

Activations Density 0.009%

words related to the acronym 'UD'

references to "UD" and related terms that denote a specific entity or concept

No Comments

No Known Activations