INDEX

Explanations

Mixed languages and code

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_32.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.32.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

æ§½

-0.27

fine

-0.27

illac

-0.26

Pel

-0.26

FIN

-0.25

ollar

-0.25

ç»Ĩ

-0.25

ilateral

-0.25

Sin

-0.25

 Finch

-0.24

POSITIVE LOGITS

ThemeProvider

0.30

_que

0.28

Ko

0.26

åĲ¸æĶ¶

0.25

enic

0.25

endez

0.25

ä¸ĢåĪĨ

0.25

Verifier

0.25

 queue

0.24

 Hermes

0.24

Activations Density 0.008%

Mixed languages and code

No Known Activations