INDEX

Explanations

scientific and medical papers

np_max-act · gemini-2.0-flash

The neuron selectively lights up on long, multisyllabic, technical/scientific terms (especially those with Greek/Latin roots or formal suffixes).

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

น

-0.07

(inv

-0.07

-0.06

Liz

-0.06

/API

-0.06

 Madame

-0.06

 Optical

-0.06

MMM

-0.06

 Mach

-0.06

볼

-0.06

POSITIVE LOGITS

 residual

0.07

idual

0.07

 permanent

0.07

 JComboBox

0.07

 протягом

0.06

pons

0.06

 Connie

0.06

ologic

0.06

 Recomm

0.06

urator

0.06

Activations Density 0.396%

scientific and medical papers

The neuron selectively lights up on long, multisyllabic, technical/scientific terms (especially those with Greek/Latin roots or formal suffixes).

No Comments

No Known Activations