INDEX

Explanations

Tests/Quizzes

np_max-act · gemini-2.0-flash

The neuron flags instructional prompt language—words and phrases that introduce or direct the user to answer, choose, or write (i.e. question‐and‐answer directives).

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

ynthesis

-0.08

職

-0.07

brities

-0.06

 hospitalized

-0.06

 parses

-0.06

 Move

-0.06

 conform

-0.06

看着

-0.06

textContent

-0.06

けて

-0.06

POSITIVE LOGITS

drm

0.07

PLY

0.07

 Maximum

0.07

 Carb

0.07

_MODAL

0.07

 optimum

0.07

anted

0.06

ενο

0.06

 Butt

0.06

 interception

0.06

Activations Density 0.016%

Tests/Quizzes

The neuron flags instructional prompt language—words and phrases that introduce or direct the user to answer, choose, or write (i.e. question‐and‐answer directives).

No Comments

No Known Activations