INDEX

Explanations

in

np_max-act · gemini-2.0-flash

The neuron activates on sentences stating that one mathematical object “is” related to another (e.g. “L₁ is parallel to L₂”), i.e. on definitional “is–to” relation phrases.

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

 erotische

-0.07

 ratio

-0.06

 situace

-0.06

桥

-0.06

 actividad

-0.06

 outskirts

-0.06

 propagation

-0.06

Properties

-0.06

 sensations

-0.06

 Erect

-0.06

POSITIVE LOGITS

_SIDE

0.08

那个

0.07

ισμ

0.07

 اعلام

0.06

 banker

0.06

 ],↵

0.06

张

0.06

 hangi

0.06

 символ

0.06

ों,

0.06

Activations Density 0.012%

in

The neuron activates on sentences stating that one mathematical object “is” related to another (e.g. “L₁ is parallel to L₂”), i.e. on definitional “is–to” relation phrases.

No Comments

No Known Activations

in

The neuron activates on sentences stating that one mathematical object “is” related to another (e.g. “L₁ is parallel to L₂”), i.e. on definitional “is–to” relation phrases.

No Comments

No Known Activations