INDEX

Explanations

Code

np_max-act · gemini-2.0-flash

The neuron consistently activates on tokens that contain non-standard characters—especially digits or non-ASCII/accented letters—i.e. markers of numbers or accented text.

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

ITTLE

-0.09

anguage

-0.08

inux

-0.07

ousy

-0.07

 인정

-0.07

 tedavi

-0.06

 філ

-0.06

aunch

-0.06

	friend

-0.06

gebra

-0.06

POSITIVE LOGITS

Latch

0.09

luk

0.09

 Lamar

0.09

LJ

0.08

 lemon

0.08

 Lester

0.08

 Lucas

0.08

LM

0.08

Lal

0.08

LC

0.08

Activations Density 0.652%

Code

The neuron consistently activates on tokens that contain non-standard characters—especially digits or non-ASCII/accented letters—i.e. markers of numbers or accented text.

No Comments

No Known Activations