INDEX

Explanations

Question answering

np_max-act · gemini-2.0-flash

The neuron flags the instructional “fill‐in‐the‐blank” style prompt language (e.g. tokens in phrases like “Fill in the following sentence,” “blank,” “sentence,” etc.).

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

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

 datingsider

-0.07

_el

-0.07

耀

-0.07

一个人

-0.06

_digits

-0.06

皇

-0.06

änner

-0.06

 อย

-0.06

 أما

-0.06

Що

-0.06

POSITIVE LOGITS

ETHOD

0.07

/interfaces

0.07

-def

0.06

.freq

0.06

 stmt

0.06

blas

0.06

 '/')↵

0.06

 พระ

0.06

anceled

0.06

 unequiv

0.06

Activations Density 0.027%

Question answering

The neuron flags the instructional “fill‐in‐the‐blank” style prompt language (e.g. tokens in phrases like “Fill in the following sentence,” “blank,” “sentence,” etc.).

No Comments

No Known Activations