INDEX

Explanations

numbers related to percentages

oai_token-act-pair · gpt-3.5-turbo

significant numerical or quantitative information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Allison

-0.99

 Alison

-0.91

 Hussain

-0.90

ahon

-0.89

-0.87

 Ludwig

-0.85

 condem

-0.85

 Henri

-0.84

Erd

-0.84

POSITIVE LOGITS

 Spur

1.47

Sp

1.43

sp

1.34

SP

1.27

 Span

1.26

SP

1.25

Sp

1.22

 Spike

1.21

Cap

1.18

 Spac

1.18

Activations Density 0.392%

numbers related to percentages

significant numerical or quantitative information

No Comments

No Known Activations