INDEX

Explanations

non-English characters and words

oai_token-act-pair · gpt-3.5-turbo

characters or symbols related to non-Latin scripts or encoding formats

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 GOODMAN

-0.91

ierrez

-0.82

emonium

-0.78

combe

-0.73

assador

-0.71

 Clover

-0.71

arton

-0.70

onomy

-0.69

 eleph

-0.68

cliffe

-0.67

POSITIVE LOGITS

Ñ

1.44

Ð½

1.42

ÑĤ

1.39

Ð´

1.36

Ðº

1.35

Ñı

1.35

Ð

1.34

ÑĢ

1.32

Ð¼

1.28

Ðµ

1.25

Activations Density 0.014%

non-English characters and words

characters or symbols related to non-Latin scripts or encoding formats

No Known Activations