INDEX

Explanations

Location contexts

np_max-act · gemini-2.0-flash

This neuron is selectively activated by the Russian word “здесь” (i.e. “here”), especially when it appears as a discourse‐opening location marker.

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

약

-0.06

	method

-0.06

ّل

-0.06

	reload

-0.06

scheme

-0.06

ंगल

-0.06

 πάνω

-0.06

ерж

-0.06

 Players

-0.06

POSITIVE LOGITS

●

0.07

 where

0.07

�

0.07

urent

0.07

seud

0.07

 Suriye

0.06

soles

0.06

álně

0.06

enuine

0.06

 organizer

0.06

Activations Density 0.049%

Location contexts

This neuron is selectively activated by the Russian word “здесь” (i.e. “here”), especially when it appears as a discourse‐opening location marker.

No Comments

No Known Activations