© 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 · 19-RESID-POST-AA · 70673 ｜ Neuronpedia

Home
Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
Qwen2.5-7B-IT
Resid Post - 131k
19-RESID-POST-AA
70673

INDEX

Explanations

repeat

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

Not in Any Lists

No Comments

Negative Logits

 disqualified

-0.07

 worsening

-0.07

 fathers

-0.07

 THESE

-0.07

 Prem

-0.07

first

-0.06

-0.06

uating

-0.06

Null

-0.06

rtl

-0.06

POSITIVE LOGITS

RelativeTo

0.08

玠

0.07

 حين

0.07

 המת

0.07

按摩

0.07

[self

0.07

 בלי

0.07

_TRACK

0.07

cfg

0.07

ayı

0.07

Activations Density 0.006%

No Known Activations