INDEX

Explanations

words related to negative human traits such as greed and deceit

oai_token-act-pair · gpt-3.5-turbo

words associated with greed and its detrimental effects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

nom

-0.76

Medic

-0.72

Marginal

-0.72

ushi

-0.71

uments

-0.70

 Parliamentary

-0.68

 Serge

-0.67

Ru

-0.67

udeau

-0.67

 Defence

-0.65

POSITIVE LOGITS

 greed

1.65

idget

1.64

laz

1.42

 craz

1.25

 carn

1.25

Laz

1.19

reck

1.12

 fastball

1.07

 Crowley

1.05

 greedy

1.01

Activations Density 0.063%

words related to negative human traits such as greed and deceit

words associated with greed and its detrimental effects

No Comments

No Known Activations