INDEX

Explanations

formulations discussing reduction or being reduced

oai_token-act-pair · gpt-3.5-turbo

mentions of decreased or diminished values or quantities

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

ï¸ı

-0.78

ahoo

-0.70

anky

-0.70

faith

-0.69

istry

-0.68

earth

-0.67

 Ange

-0.66

stars

-0.65

ive

-0.64

amera

-0.63

POSITIVE LOGITS

 reduced

3.34

 Reduced

2.35

 decreased

2.19

 lowered

2.06

 diminished

2.02

 increased

1.94

 reduction

1.85

 shortened

1.84

 reductions

1.83

 reducing

1.61

Activations Density 0.017%

formulations discussing reduction or being reduced

mentions of decreased or diminished values or quantities

No Comments

No Known Activations