INDEX

Explanations

expressions of understanding and comprehension

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

"understand" or "see" followed by a justification

np_acts-logits-general · gemini-2.0-flash

understanding why

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 24-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.24.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

 Füße

-0.42

 Verfahrens

-0.40

 świad

-0.38

 sabido

-0.38

 DISEASES

-0.38

 חיצוניים

-0.38

empted

-0.37

 enfermed

-0.37

 picioare

-0.36

 odbior

-0.35

POSITIVE LOGITS

why

0.84

why

0.67

WHY

0.61

WHY

0.58

Why

0.58

Why

0.57

 nahilalakip

0.54

 CallOverrides

0.54

 understand

0.53

 المعيارى

0.53

Activations Density 0.223%

expressions of understanding and comprehension

"understand" or "see" followed by a justification

understanding why

No Comments

No Known Activations

expressions of understanding and comprehension

"understand" or "see" followed by a justification

understanding why

No Comments

No Known Activations