© 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 · 19-RESID-POST-AA · 118562 ｜ Neuronpedia

Home
Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
Llama3.1-8B-IT (Instruct)
Resid Post - 131k
19-RESID-POST-AA
118562

INDEX

Explanations

which

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_19/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.19.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

:'.$

-0.07

agrant

-0.07

	GameObject

-0.07

Don

-0.07

Don

-0.07

 Saturday

-0.06

-0.06

Say

-0.06

 ReSharper

-0.06

-0.06

POSITIVE LOGITS

 which

0.17

which

0.14

 WHICH

0.11

 Which

0.10

—which

0.09

Which

0.08

 whilst

0.08

(which

0.07

 koje

0.07

 والتي

0.07

Activations Density 0.170%

No Known Activations