INDEX

Explanations

understand

np_max-act · gemini-2.0-flash

The neuron fires on the model’s self-descriptions of its language capabilities—especially verbs like “understand/understanding” and “generate/generating” when talking about what it can do.

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

Negative Logits

альных

-0.08

 нічого

-0.07

تور

-0.07

 shipped

-0.07

 méně

-0.06

	vertices

-0.06

оль

-0.06

 مرگ

-0.06

 новый

-0.06

	exit

-0.06

POSITIVE LOGITS

owering

0.07

 ullam

0.07

loid

0.06

/Runtime

0.06

ymes

0.06

 transl

0.06

 vysok

0.06

दर

0.06

polate

0.06

org

0.06

Activations Density 0.010%

understand

The neuron fires on the model’s self-descriptions of its language capabilities—especially verbs like “understand/understanding” and “generate/generating” when talking about what it can do.

No Known Activations

understand

The neuron fires on the model’s self-descriptions of its language capabilities—especially verbs like “understand/understanding” and “generate/generating” when talking about what it can do.

No Known Activations