INDEX

Explanations

a

np_max-act · gemini-2.0-flash

code snippets related to programming in Python.

oai_token-act-pair · gpt-4o-mini Triggered by @xinyanhu8

The neuron is principally detecting the indefinite articles “a” and “an.”

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

Detects imperative user requests asking the assistant to create or produce something (commands to make/build/etc.).

oai_token-act-pair · gpt-5-mini Triggered by @vetterc0

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

โจ

-0.07

 عضو

-0.07

\Field

-0.06

菜

-0.06

variably

-0.06

_agents

-0.06

icycle

-0.06

.Users

-0.06

\Tests

-0.06

=torch

-0.06

POSITIVE LOGITS

\v

0.07

.Av

0.07

\e

0.06

_COMPONENT

0.06

 الن

0.06

_EXP

0.06

_SUS

0.06

abric

0.06

 callbacks

0.06

_arm

0.06

Activations Density 0.053%

a

code snippets related to programming in Python.

The neuron is principally detecting the indefinite articles “a” and “an.”

Detects imperative user requests asking the assistant to create or produce something (commands to make/build/etc.).

No Comments

No Known Activations