© Neuronpedia 2026

Privacy & Terms Blog GitHub Slack Twitter Contact

Jacobian LensNEW

Natural Language

MAY 26 Assistant AxisJAN 26 Circuit TracerAPR 26Steer SAE Evals Exports interp-engineNEW Community Blog Privacy & Terms Contact

QWEN2.5-7B-IT · 19-RESID-POST-AA · 113086 ｜ Neuronpedia

Home
Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
Qwen2.5-7B-IT
Resid Post - 131k
19-RESID-POST-AA
113086

INDEX

Explanations

SQL code

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Top Features by Cosine Similarity

Configuration

andyrdt/saes-qwen2.5-7b-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

Negative Logits

_Act

-0.07

 conservatives

-0.07

ror

-0.07

Creating

-0.07

道士

-0.06

.EntityFrameworkCore

-0.06

_Project

-0.06

專

-0.06

 secondo

-0.06

».

-0.06

POSITIVE LOGITS

咛

0.08

IntArray

0.07

bite

0.07

cents

0.07

Summary

0.07

PNG

0.07

alignment

0.07

(Link

0.07

 milan

0.06

恙

0.06

Activations Density 0.010%

No Known Activations