INDEX

Explanations

phrases related to assistance and user support

oai_token-act-pair · gpt-4o-mini Triggered by @bot

second person actions and possessions

np_acts-logits-general · gemini-2.5-flash-lite

The neuron detects second-person references — words and verb forms addressing "you" (e.g., English "you" and equivalent Spanish/Portuguese second-person forms).

oai_token-act-pair · gpt-5-mini Triggered by @vetterc0

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 بتوان

-0.43

 themselves

-0.40

 their

-0.39

 می‌توان

-0.36

 they

-0.35

overset

-0.34

dis

-0.33

re

-0.33

他自己

-0.32

ensement

-0.32

POSITIVE LOGITS

 yourself

1.23

yourself

1.13

your

1.08

 yourselves

1.08

 your

1.05

 Yourself

0.96

Your

0.94

Yourself

0.93

 YOURSELF

0.91

あなたが

0.89

Activations Density 1.694%

phrases related to assistance and user support

second person actions and possessions

The neuron detects second-person references — words and verb forms addressing "you" (e.g., English "you" and equivalent Spanish/Portuguese second-person forms).

No Comments

No Known Activations