INDEX

Explanations

phrases related to lists and rankings

oai_token-act-pair · gpt-3.5-turbo

punctuation marks, particularly periods and question marks

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 credential

-0.81

 nurt

-0.74

 emanc

-0.73

 contextual

-0.73

 unle

-0.71

 spatial

-0.70

 authorized

-0.70

âĢĳ

-0.70

 enrol

-0.69

 safegu

-0.68

POSITIVE LOGITS

 Anyway

1.55

 Seems

1.55

 Personally

1.37

 Also

1.37

 Maybe

1.35

 Honestly

1.33

:(

1.33

Anyway

1.31

 Hopefully

1.31

 Basically

1.30

Activations Density 0.445%

phrases related to lists and rankings

punctuation marks, particularly periods and question marks

No Comments

No Known Activations