© 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 · 119910 ｜ Neuronpedia

Home
Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
Qwen2.5-7B-IT
Resid Post - 131k
15-RESID-POST-AA
119910

INDEX

Explanations

search

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

",__

-0.08

琪

-0.07

薯

-0.07

 øns

-0.07

צי

-0.07

搡

-0.06

atie

-0.06

-transition

-0.06

 nghi

-0.06

生态圈

-0.06

POSITIVE LOGITS

 remain

0.08

 Dogs

0.07

_probs

0.07

[mid

0.07

 Extreme

0.06

Ere

0.06

 Sudoku

0.06

omaly

0.06

 extreme

0.06

un

0.06

Activations Density 0.014%

No Known Activations