INDEX

Explanations

source code

np_max-act-logits · gemini-2.0-flash

source code artifacts like pragmas, headers, comments, and version information.

oai_token-act-pair · claude-3-5-haiku-20241022 Triggered by @kdearsty

legal terms

np_max-act · claude-3-5-haiku-20241022 Triggered by @kdearsty

say programming syntax

np_max-act-logits · claude-3-5-haiku-20241022 Triggered by @kdearsty

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_27.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.27.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ð´Ð²Ð¸Ð³Ð°

-0.28

ington

-0.27

æ°Ľ

-0.26

INGTON

-0.26

igung

-0.25

å¾Īæĥ³

-0.24

æĦĨ

-0.24

åī§

-0.24

 trá»Ŀi

-0.24

æĿĥ

-0.23

POSITIVE LOGITS

fk

0.27

FK

0.27

icks

0.27

å¾ĹæĦı

0.26

åĩºæĿ¥

0.26

 SelectListItem

0.25

 bend

0.24

ificent

0.23

Gay

0.23

iving

0.23

Activations Density 0.884%

source code

source code artifacts like pragmas, headers, comments, and version information.

legal terms

say programming syntax

No Comments

No Known Activations