INDEX

Explanations

references to evasion and escape from danger

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

escaping or hiding

np_acts-logits-general · gemini-2.0-flash

fleeing or hiding from arrest

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 ModelExpression

-0.61

VersionUID

-0.52

UnusedPrivate

-0.51

 المعيارى

-0.51

$_"

-0.48

 viewDidLoad

-0.46

colnshire

-0.45

 Penh

-0.44

MemoryWarning

-0.44

APORE

-0.43

POSITIVE LOGITS

 hiding

0.71

逃

0.69

 fleeing

0.69

 escape

0.69

 escaping

0.68

escape

0.67

 flee

0.67

 Flucht

0.66

 hide

0.65

 dodge

0.65

Activations Density 0.378%

references to evasion and escape from danger

escaping or hiding

fleeing or hiding from arrest

No Comments

No Known Activations