INDEX

Explanations

years and historical periods

oai_token-act-pair · gpt-3.5-turbo

sentences or clauses that are declarative in nature

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 sprint

-0.89

 selfie

-0.82

 purse

-0.82

 jersey

-0.82

 emergency

-0.81

 pocket

-0.80

plet

-0.79

 emoji

-0.79

 cough

-0.78

 vape

-0.78

POSITIVE LOGITS

 Eventually

1.23

 Known

1.22

 Although

1.22

 Throughout

1.20

 However

1.20

 Originally

1.19

 Ironically

1.19

 Nevertheless

1.18

 During

1.17

 Later

1.17

Activations Density 0.491%

years and historical periods

sentences or clauses that are declarative in nature

No Comments

No Known Activations