INDEX

Explanations

phrases that indicate recent developments or updates in various contexts

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

recent past or completion

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

theoremstyle

-0.46

picasso

-0.44

səhifə

-0.44

lably

-0.43

ématiques

-0.42

 gewöhnlich

-0.41

UnsafeEnabled

-0.41

oredCriteria

-0.41

eriam

-0.41

 Ruin

-0.41

POSITIVE LOGITS

Aholisi

0.47

Recently

0.40

 Recent

0.40

 typelib

0.40

Current

0.39

 prochains

0.37

 getItemCount

0.37

cribir

0.37

 Recently

0.37

Recent

0.37

Activations Density 0.030%

phrases that indicate recent developments or updates in various contexts

recent past or completion

No Comments

No Known Activations