INDEX

Explanations

mentions of personal characteristics or traits

oai_token-act-pair · gpt-3.5-turbo

mentions of specific individuals and their characteristics or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

 Millennium

-0.79

 Illegal

-0.67

 Mandatory

-0.63

 Replacement

-0.63

Prem

-0.63

 Voltage

-0.62

ãĥķãĤ¡

-0.62

 Probe

-0.62

rame

-0.60

FTWARE

-0.60

POSITIVE LOGITS

 prefers

1.56

 enjoys

1.36

 tends

1.34

 loves

1.28

 strives

1.26

 possesses

1.26

 avoids

1.23

 rarely

1.22

 seldom

1.22

 hates

1.21

Activations Density 0.345%

mentions of personal characteristics or traits

mentions of specific individuals and their characteristics or actions

No Comments

No Known Activations