© Neuronpedia 2026

Privacy & Terms Blog GitHub Slack Twitter Contact

Neuronpedia

APIAssistant AxisNEW Circuit TracerNEW Steer SAE Evals Exports Community Blog Privacy & Terms Contact

LLAMA3.1-8B-IT · 11-RESID-POST-AA · 88248 ｜ Neuronpedia

Home
Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
Llama3.1-8B-IT (Instruct)
Resid Post - 131k
11-RESID-POST-AA
88248

INDEX

Explanations

uphold

np_max-act · gemini-2.0-flash

New Auto-Interp

Top Features by Cosine Similarity

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

Show PlotsShow ExplanationShow ActivationsShow Test FieldShow SteerShow Link

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 disagreement

-0.07

 easy

-0.07

 penis

-0.07

 TERMS

-0.07

Fat

-0.07

 desires

-0.07

	lines

-0.07

ategies

-0.07

azi

-0.06

cliente

-0.06

POSITIVE LOGITS

 uphold

0.09

 upheld

0.08

扬

0.07

luğ

0.06

 wykon

0.06

履

0.06

.pageX

0.06

 sustained

0.06

 fostering

0.06

eated

0.06

Activations Density 0.002%

No Known Activations