INDEX

Explanations

terms related to physical parts, components, or elements, often in a manufacturing or technical context

oai_token-act-pair · gpt-3.5-turbo

variations of the suffix "ings," indicative of processes or actions

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

 earthqu

-0.82

vier

-0.79

SIGN

-0.76

PDATE

-0.62

ISON

-0.62

Ub

-0.62

URRENT

-0.61

ACP

-0.61

ãĥ¯ãĥ³

-0.61

BIL

-0.61

POSITIVE LOGITS

hots

1.28

omething

1.22

tons

1.10

creen

1.07

pace

1.01

hot

1.00

cape

0.99

peed

0.94

ystem

0.92

hower

0.92

Activations Density 0.062%

terms related to physical parts, components, or elements, often in a manufacturing or technical context

variations of the suffix "ings," indicative of processes or actions

No Comments

No Known Activations