INDEX

Explanations

person in charge

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

Not in Any Lists

No Comments

Negative Logits

åĿĽ

-0.38

å¾Ĺ

-0.28

éĹ»

-0.26

empt

-0.26

 tempt

-0.25

çº·

-0.25

ä¹į

-0.25

 havoc

-0.25

æĹłæīĢè°ĵ

-0.25

å¯»

-0.25

POSITIVE LOGITS

 controlled

0.44

 managed

0.41

 operated

0.39

 coached

0.36

 dedicated

0.36

 funded

0.35

 owned

0.35

 guided

0.35

 anchored

0.34

 held

0.34

Activations Density 0.031%

person in charge

No Comments

No Known Activations