INDEX

Explanations

proper nouns or names containing 'Ng' or 'Ig'

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Ng" and related variants

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ãĥĩãĤ£

-0.85

fare

-0.79

Madison

-0.77

-0.75

Catal

-0.72

furt

-0.70

bre

-0.70

BRE

-0.69

ãĤ·ãĥ£

-0.69

ãĥĵ

-0.68

POSITIVE LOGITS

Ng

1.38

ongs

0.89

ocom

0.88

atana

0.87

yang

0.87

achu

0.85

wana

0.84

ong

0.84

aleb

0.84

obi

0.83

Activations Density 0.004%

proper nouns or names containing 'Ng' or 'Ig'

mentions of the name "Ng" and related variants

No Comments

No Known Activations