© Neuronpedia 2026

Privacy & Terms Blog GitHub Slack Twitter Contact

Neuronpedia

Natural Language

NEW Assistant AxisNEW Circuit TracerUPDATESteer SAE Evals ExportsAPI Community Blog Privacy & Terms Contact

LLAMA3.1-8B-IT · 27-RESID-POST-AA · 45945 ｜ Neuronpedia

Home
Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
Llama3.1-8B-IT
Resid Post - 131k
27-RESID-POST-AA
45945

INDEX

Explanations

guard

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Top Features by Cosine Similarity

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_27/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.27.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

Show PlotsShow ExplanationShow ActivationsShow Test FieldShow SteerShow Link

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 peptide

-0.07

(title

-0.07

 ontology

-0.07

ватися

-0.07

 Bucket

-0.07

 writers

-0.07

property

-0.06

 Alice

-0.06

 Elizabeth

-0.06

Frequency

-0.06

POSITIVE LOGITS

 Guard

0.10

0.09

 guard

0.09

_Pr

0.08

-guard

0.08

Guard

0.08

0.07

 Guardians

0.07

agr

0.07

0.07

Activations Density 0.007%

No Known Activations