INDEX

Explanations

on

np_max-act · gemini-2.0-flash

tokens that appear in structured metadata or form-like fields (labels, registration/date fields, IDs and similar technical/form labels).

oai_token-act-pair · gpt-5-mini Triggered by @vetterc0

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

，她

-0.07

rina

-0.06

 cartesian

-0.06

.setOnClickListener

-0.06

scan

-0.06

//----------------------------------------------------------------------------------------------------------------

-0.06

 cones

-0.06

�

-0.06

 Banana

-0.06

POSITIVE LOGITS

πει

0.07

만

0.07

Along

0.07

 فروش

0.06

 đặc

0.06

alloc

0.06

wechat

0.06

IMITIVE

0.06

ului

0.06

White

0.06

Activations Density 7.976%

on

tokens that appear in structured metadata or form-like fields (labels, registration/date fields, IDs and similar technical/form labels).

No Comments

No Known Activations