INDEX

Explanations

code and math

np_max-act · gemini-2.0-flash

code syntax

np_max-act · deepseek-v3 Triggered by @finnwit

Token "in" to describe location of an abstract object, mostly used in pointing to a location in a symbolic programming structure like a controller or window, otherwise functions to describe inclusion.

eleuther_acts_top20 · deepseek-v3 Triggered by @finnwit

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_20.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.20.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

amble

-0.29

ä¸Ģå¤ľ

-0.26

çķ¥

-0.26

æ½¢

-0.26

çķ¥å¾®

-0.26

lator

-0.26

ç¨įå¾®

-0.25

bomb

-0.25

allen

-0.25

 Bomb

-0.24

POSITIVE LOGITS

lak

0.29

æī§æķĻ

0.26

åłĨ

0.26

 inet

0.25

nic

0.25

åıĳçİ°

0.24

lf

0.24

 crÃ©d

0.24

eas

0.23

ãģªãĤĤãģ®

0.23

Activations Density 11.510%

code and math

code syntax

Token "in" to describe location of an abstract object, mostly used in pointing to a location in a symbolic programming structure like a controller or window, otherwise functions to describe inclusion.

No Comments

No Known Activations

code and math

code syntax

Token "in" to describe location of an abstract object, mostly used in pointing to a location in a symbolic programming structure like a controller or window, otherwise functions to describe inclusion.

No Comments

No Known Activations