INDEX

Explanations

that

np_max-act · gemini-2.0-flash

This neuron detects the “do anything that …” style instruction phrases that attempt to override the model’s normal rules.

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rawer

-0.07

文字

-0.07

 helicopters

-0.06

BYTE

-0.06

HB

-0.06

 Amber

-0.06

adiator

-0.06

ネル

-0.06

reshold

-0.06

htaking

-0.06

POSITIVE LOGITS

 Scala

0.07

사람

0.06

……↵↵

0.06

 wirklich

0.06

θούν

0.06

!!!!↵↵

0.06

اما

0.06

$↵

0.06

 tehdy

0.06

 interior

0.06

Activations Density 0.003%

that

This neuron detects the “do anything that …” style instruction phrases that attempt to override the model’s normal rules.

No Comments

No Known Activations