INDEX

Explanations

instances of the word "something"

oai_token-act-pair · gpt-3.5-turbo

references to "something" indicating a desire for change or new experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

oret

-0.70

inders

-0.66

anos

-0.65

 Chiefs

-0.64

ortex

-0.64

 Examiner

-0.64

assis

-0.64

awar

-0.63

DOM

-0.63

nor

-0.62

POSITIVE LOGITS

Else

1.46

 else

1.38

 resembling

0.95

 Else

0.88

 worthwhile

0.85

 intangible

0.83

else

0.80

 nutrit

0.79

 innocuous

0.77

 incomprehensible

0.77

Activations Density 0.036%

instances of the word "something"

references to "something" indicating a desire for change or new experiences

No Comments

No Known Activations