INDEX

Explanations

negative descriptions or critiques related to intelligence or decision-making

oai_token-act-pair · gpt-3.5-turbo

instances of the word "dumb" and variations thereof

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

iday

-0.92

icol

-0.71

icum

-0.67

 Seek

-0.66

 Metropolitan

-0.65

 Mara

-0.65

 Advisory

-0.65

 Chrys

-0.64

eria

-0.64

seek

-0.64

POSITIVE LOGITS

 dumb

3.62

 Dumb

2.20

 stupid

2.09

 Stupid

1.50

 retarded

1.43

 stupidity

1.42

 foolish

1.42

 lame

1.40

 ignorant

1.40

 silly

1.38

Activations Density 0.008%

negative descriptions or critiques related to intelligence or decision-making

instances of the word "dumb" and variations thereof

No Comments

No Known Activations