INDEX

Explanations

Greek letters and symbols

oai_token-act-pair · gpt-3.5-turbo

scientific terms and concepts related to ancient Greek and Roman studies

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

Wim

-0.85

hower

-0.82

 Camel

-0.74

 Forest

-0.74

 Assass

-0.72

Mer

-0.72

 Mush

-0.71

 Georgetown

-0.70

 Tiger

-0.70

 declass

-0.70

POSITIVE LOGITS

Ïģ

3.20

Î±

3.10

Î½

3.09

Ï

3.01

Îº

2.98

Î»

2.94

Î³

2.93

Ïĥ

2.91

Î¿

2.90

ÏĦ

2.87

Activations Density 0.024%

Greek letters and symbols

scientific terms and concepts related to ancient Greek and Roman studies

No Comments

No Known Activations