INDEX

Explanations

Le

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_30.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.30.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

å¯¼

-0.27

 newName

-0.26

æ¸¸

-0.26

æ½¤

-0.25

TT

-0.24

å¹²

-0.24

è´°

-0.23

rad

-0.23

æĳ¸

-0.23

æĽ´å¥½çļĦ

-0.23

POSITIVE LOGITS

ÑĢÐ°Ð¶Ð°

0.26

Ĳľ

0.26

otten

0.25

 fidelity

0.25

ç¼Ļ

0.24

 ÑģÐµÐ±Ðµ

0.23

Ã¤ge

0.23

-loading

0.23

ÑĥÐº

0.23

SKIP

0.23

Activations Density 0.002%

Le

No Comments

No Known Activations