INDEX

Explanations

Proper nouns

np_max-act · gemini-2.0-flash

capitalized proper names, especially surnames and eponymous terms, appearing in technical or news text.

oai_token-act-pair · gpt-5 Triggered by @liuliujiujiu05

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_7.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.7.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

çļĦæĪĲåĬŁ

-0.29

Ð»Ð°Ð´

-0.28

æĪĲåĬŁ

-0.27

 rightfully

-0.26

çĭŃ

-0.26

æľ¬æľŁ

-0.25

ç»Īäºİ

-0.25

 Wander

-0.25

å¼Ģåı£

-0.25

.success

-0.25

POSITIVE LOGITS

PECT

0.29

å¿łå®ŀ

0.29

 kiá»ĩn

0.25

 univers

0.25

sal

0.24

aN

0.24

ä¸ªå¤ļ

0.24

éĴ¤

0.23

mut

0.23

pg

0.23

Activations Density 0.001%

Proper nouns

capitalized proper names, especially surnames and eponymous terms, appearing in technical or news text.

No Comments

No Known Activations