INDEX

Explanations

"linear" patterns or concepts in text

oai_token-act-pair · gpt-3.5-turbo

terminology related to linear concepts and relationships

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

enda

-0.79

abies

-0.78

 Bills

-0.77

 Heard

-0.74

acca

-0.68

orthy

-0.67

 Checks

-0.67

 canv

-0.66

assador

-0.66

yd

-0.65

POSITIVE LOGITS

linear

3.43

 linear

3.24

 Linear

2.26

 sequential

1.40

Lin

1.26

 exponential

1.25

 regression

1.21

onential

1.05

 radial

1.05

 equilibrium

1.03

Activations Density 0.039%

"linear" patterns or concepts in text

terminology related to linear concepts and relationships

No Comments

No Known Activations