INDEX

Explanations

references to best practices

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

best practice

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

 Christmas

-0.36

arrayOf

-0.32

 statues

-0.32

quila

-0.31

 vast

-0.30

кина

-0.30

 arrayOf

-0.30

 claimed

-0.30

 compound

-0.29

 jSON

-0.29

POSITIVE LOGITS

 practices

0.84

OGND

0.74

practices

0.73

 Practices

0.73

 nakalista

0.69

Practices

0.67

 prácticas

0.67

 recomendaciones

0.66

 estándares

0.66

Datuak

0.66

Activations Density 0.006%

references to best practices

best practice

No Comments

No Known Activations