© Neuronpedia 2026

Privacy & Terms Blog GitHub Slack Twitter Contact

Neuronpedia

Jacobian LensNEW

Natural Language

NEW Assistant AxisNEW Circuit TracerUPDATESteer SAE Evals Exports interp-engineNEWAPI Community Blog Privacy & Terms Contact

QWEN2.5-7B-IT · 15-RESID-POST-AA · 51061 ｜ Neuronpedia

Home
Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
Qwen2.5-7B-IT
Resid Post - 131k
15-RESID-POST-AA
51061

INDEX

Explanations

file paths and code

np_max-act · gemini-2.0-flash

New Auto-Interp

Top Features by Cosine Similarity

Configuration

andyrdt/saes-qwen2.5-7b-instruct/resid_post_layer_15/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.15.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

婉

-0.07

孟

-0.07

 millennium

-0.07

箔

-0.07

-0.07

Տ

-0.07

ǔ

-0.06

 trend

-0.06

bam

-0.06

=B

-0.06

POSITIVE LOGITS

 CRUD

0.07

 Parents

0.07

휀

0.07

 internally

0.07

ubern

0.07

ETING

0.07

 Anaheim

0.07

 Ethernet

0.07

 EditorGUI

0.06

تحد

0.06

Activations Density 0.010%

No Known Activations