INDEX

Explanations

phrases related to features, benefits, and enhancements

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing the variety and abundance of options or features

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

rison

-0.70

ichen

-0.63

 Lumpur

-0.62

glas

-0.60

lash

-0.58

 Aware

-0.57

 WARN

-0.56

 Borders

-0.54

 Correct

-0.54

olitan

-0.53

POSITIVE LOGITS

 assorted

1.13

etc

1.09

 downright

1.05

 myriad

1.00

anything

0.99

cellaneous

0.93

 countless

0.85

everything

0.84

 Lastly

0.84

 endless

0.84

Activations Density 0.226%

phrases related to features, benefits, and enhancements

phrases emphasizing the variety and abundance of options or features

No Comments

No Known Activations