INDEX

Explanations

comma

np_max-act · gemini-2.0-flash

This neuron detects explicit content descriptors—words that label or trigger vulgar, offensive, hateful, perverted, violent, or otherwise disallowed/immoral material.

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

omet

-0.07

Roman

-0.06

arrings

-0.06

InOut

-0.06

 Destiny

-0.06

 pattern

-0.06

NTAX

-0.06

ちら

-0.06

い

-0.06

POSITIVE LOGITS

 Đào

0.06

�

0.06

 poru

0.06

.getUserId

0.06

 раньше

0.06

_ping

0.06

 São

0.06

 лак

0.06

 submar

0.06

 زي

0.06

Activations Density 0.018%

comma

This neuron detects explicit content descriptors—words that label or trigger vulgar, offensive, hateful, perverted, violent, or otherwise disallowed/immoral material.

No Known Activations

comma

This neuron detects explicit content descriptors—words that label or trigger vulgar, offensive, hateful, perverted, violent, or otherwise disallowed/immoral material.

No Known Activations