INDEX

Explanations

negative descriptions of individuals or groups

oai_token-act-pair · gpt-3.5-turbo

negative characterizations of individuals or groups, particularly in a political context

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

aughtered

-0.81

ortality

-0.79

rimination

-0.77

 GOODMAN

-0.77

vation

-0.76

":"/

-0.75

estones

-0.75

isSpecialOrderable

-0.74

ITNESS

-0.74

onds

-0.72

POSITIVE LOGITS

 arrogant

1.51

 incompetent

1.40

 ignorant

1.32

 selfish

1.31

 narcissistic

1.30

 immature

1.29

 insensitive

1.29

 dishonest

1.29

 hypocritical

1.28

 lazy

1.27

Activations Density 0.265%

negative descriptions of individuals or groups

negative characterizations of individuals or groups, particularly in a political context

No Comments

No Known Activations