INDEX

Explanations

dates in a specific format

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ortion

-0.73

ãĥ¼ãĥĨãĤ£

-0.72

ãģĨ

-0.71

ï¿½ï¿½ï¿½ï¿½

-0.70

square

-0.66

acca

-0.64

JR

-0.63

ividual

-0.63

 Coulter

-0.62

CPC

-0.62

POSITIVE LOGITS

isure

1.33

ather

1.22

opard

1.22

agues

1.21

aving

1.18

aky

1.12

aping

1.11

ukemia

1.11

icester

1.09

IGH

1.06

Activations Density 0.029%

dates in a specific format

No Comments

No Known Activations