INDEX

Explanations

words related to historical figures and royalty

oai_token-act-pair · gpt-3.5-turbo

references to monarchs or royal titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

pots

-0.80

zers

-0.69

maxwell

-0.67

downs

-0.64

pmwiki

-0.62

 quantities

-0.61

 define

-0.61

uate

-0.60

 attackers

-0.59

pim

-0.58

POSITIVE LOGITS

 course

1.28

sted

0.98

course

0.85

ortunately

0.85

 Alexandria

0.77

 Anaheim

0.76

 Course

0.74

 Norwich

0.72

 ours

0.72

Ń·

0.72

Activations Density 0.137%

words related to historical figures and royalty

references to monarchs or royal titles

No Comments

No Known Activations