INDEX

Explanations

code and data snippets

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_25.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.25.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

rec

-0.27

 although

-0.26

 though

-0.26

Rec

-0.26

ÐµÐ¹ÑģÑĤÐ²

-0.25

 Dickinson

-0.25

@student

-0.24

 despite

-0.24

 rarely

-0.24

UNET

-0.24

POSITIVE LOGITS

edly

0.27

åĩĨå¤ĩå¥½

0.26

boom

0.26

wow

0.25

berry

0.25

 testcase

0.24

_phrase

0.23

 teste

0.23

 Awesome

0.23

bands

0.23

Activations Density 1.090%

code and data snippets

No Known Activations