INDEX

Explanations

mentions of lists or items being on a list

oai_token-act-pair · gpt-3.5-turbo

references related to lists and rankings

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

 prevailed

-0.70

wright

-0.69

 apprehend

-0.67

nce

-0.63

 situated

-0.62

Allows

-0.62

idian

-0.61

say

-0.60

 wrest

-0.59

icz

-0.59

POSITIVE LOGITS

 list

1.92

 lists

1.47

 List

1.46

list

1.38

 blacklist

1.31

 checklist

1.29

 Lists

1.29

 LIST

1.25

LIST

1.24

List

1.23

Activations Density 0.263%

mentions of lists or items being on a list

references related to lists and rankings

No Comments

No Known Activations