INDEX

Explanations

expressions indicating alternatives or possibilities

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

igr

-0.72

ETS

-0.68

vant

-0.65

FACE

-0.64

estro

-0.64

gorithm

-0.64

"],"

-0.62

estern

-0.61

KER

-0.61

emi

-0.60

POSITIVE LOGITS

 alternatively

1.34

 indeed

1.17

 worse

1.15

 maybe

1.13

 else

1.11

 perhaps

1.09

 rather

1.07

chard

1.06

 whatever

1.01

 even

0.97

Activations Density 0.077%

expressions indicating alternatives or possibilities

No Comments

No Known Activations