INDEX

Explanations

words related to threading and computer programming

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

 Flavoring

-0.78

cer

-0.73

mosp

-0.69

oult

-0.67

oice

-0.66

cers

-0.66

kens

-0.65

issan

-0.65

ãĤº

-0.62

 Juda

-0.60

POSITIVE LOGITS

bare

1.73

thread

1.07

pool

1.03

need

1.01

 threads

1.00

 thread

1.00

ming

0.93

ing

0.85

driver

0.83

walking

0.82

Activations Density 0.024%

words related to threading and computer programming

No Comments

No Known Activations