INDEX

Explanations

based

np_max-act · gemini-2.0-flash

This neuron detects self-descriptive disclaimers about being a text-based AI (e.g., “I am a text-based AI and cannot…”).

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

λαν

-0.07

gov

-0.07

 вже

-0.07

-current

-0.07

 NullPointerException

-0.06

combination

-0.06

 mile

-0.06

 memories

-0.06

]});↵

-0.06

Ell

-0.06

POSITIVE LOGITS

窗

0.07

ファ

0.06

 chili

0.06

elsif

0.06

 Petite

0.06

 Autos

0.06

 foundational

0.06

 constr

0.06

 Auto

0.06

mozilla

0.06

Activations Density 0.004%

based

This neuron detects self-descriptive disclaimers about being a text-based AI (e.g., “I am a text-based AI and cannot…”).

No Comments

No Known Activations