INDEX

Explanations

punctuation marks

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Ribbon

-0.52

cule

-0.52

 Gateway

-0.50

 Quad

-0.50

itialized

-0.49

ulum

-0.48

 Blueprint

-0.48

 blockbuster

-0.48

 Coalition

-0.48

 acronym

-0.47

POSITIVE LOGITS

etc

1.04

etc

0.99

 lest

0.95

 thereby

0.93

 whereas

0.91

 respectively

0.88

thus

0.83

while

0.83

according

0.82

instead

0.81

Activations Density 0.603%

punctuation marks

No Comments

No Known Activations