INDEX

Explanations

names of movies or technical terms

oai_token-act-pair · gpt-3.5-turbo

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

soType

-0.51

ktop

-0.50

],"

-0.48

 GOODMAN

-0.47

`.

-0.47

 Canaver

-0.47

clusively

-0.46

 eternity

-0.45

 Contribut

-0.45

ECA

-0.45

POSITIVE LOGITS

has

1.08

 corresponds

1.07

 tends

1.06

was

1.06

is

1.05

 ensures

1.04

 seems

1.03

 appears

1.02

 translates

1.01

 resembles

1.00

Activations Density 0.919%

names of movies or technical terms

No Comments

No Known Activations