INDEX

Explanations

words related to Eastern European language characters and terms

oai_token-act-pair · gpt-3.5-turbo

characters or symbols that appear to be non-standard or corrupted text

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

 fragmentation

-0.67

 rooting

-0.67

 equival

-0.66

 structure

-0.66

 mileage

-0.64

 marrow

-0.63

 Coliseum

-0.63

 stake

-0.62

 multipl

-0.61

 disarm

-0.61

POSITIVE LOGITS

oÄŁ

1.21

nik

1.14

pai

1.12

ÑĮ

1.12

nown

1.04

ti

1.03

tun

1.02

eh

1.02

tal

0.97

ÃŃn

0.95

Activations Density 0.036%

words related to Eastern European language characters and terms

characters or symbols that appear to be non-standard or corrupted text

No Comments

No Known Activations