INDEX

Explanations

terms related to programming concepts and structures

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

policy rules or settings

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

 gynhyrchwyd

-0.35

owohl

-0.35

 Unterscheidung

-0.34

 Unterschiede

-0.32

 ladrillo

-0.32

 znal

-0.30

 egli

-0.29

 Gedichte

-0.28

ม้

-0.28

 gefällt

-0.28

POSITIVE LOGITS

CloseOperation

0.66

 ब्रेकडाउन

0.63

ỡng

0.62

AndEndTag

0.59

 незавершена

0.58



0.56

 MainAxisSize

0.56

nodoc

0.56

 autorytatywna

0.53

SpringRunner

0.53

Activations Density 0.140%

terms related to programming concepts and structures

policy rules or settings

No Comments

No Known Activations