INDEX

Explanations

sequential elements in a series

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

ordinal numbers and references to sequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

lems

-0.79

 ILCS

-0.73

ãĤ¦ãĤ¹

-0.72

etimes

-0.69

phi

-0.69

jong

-0.68

endas

-0.68

otonin

-0.68

mbuds

-0.68

atars

-0.67

POSITIVE LOGITS

 involves

1.14

 comprises

1.08

 consists

1.07

 relates

1.07

 contains

1.00

two

0.99

is

0.98

 belongs

0.98

was

0.95

 includes

0.95

Activations Density 0.197%

sequential elements in a series

ordinal numbers and references to sequences

No Comments

No Known Activations