INDEX

Explanations

Complex sentences

np_max-act · gemini-2.0-flash

The neuron strongly activates on tokens in clauses describing or endorsing “immoral” or “unethical” behavior, essentially spotting mentions of ethics violations.

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

content related to harmful, unethical, or immoral behavior.

oai_token-act-pair · gpt-4o-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

false

-0.07

Nation

-0.07

 pools

-0.06

 Forced

-0.06

宮

-0.06

 Chase

-0.06

AK

-0.06

administrator

-0.06

 software

-0.06

 Він

-0.06

POSITIVE LOGITS

(thing

0.07

 consequential

0.06

 veřejné

0.06

SSH

0.06

__',

0.06

>(

0.06

igation

0.06

groundColor

0.06

.google

0.06

=o

0.06

Activations Density 0.004%

Complex sentences

The neuron strongly activates on tokens in clauses describing or endorsing “immoral” or “unethical” behavior, essentially spotting mentions of ethics violations.

content related to harmful, unethical, or immoral behavior.

No Comments

No Known Activations