INDEX

Explanations

references to new or recent developments, such as updates or releases

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

latest news and updates

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

 myö

-0.38

 parall

-0.37

 parallel

-0.37

zelfde

-0.37

 duide

-0.36

 same

-0.35

 néer

-0.35

 pozdě

-0.35

 mesma

-0.34

 tonel

-0.34

POSITIVE LOGITS

newest

0.86

 LATEST

0.81

latest

0.81

 latest

0.78

Latest

0.76

 newest

0.75

Newest

0.75

最新的

0.72

LATEST

0.70

 Newest

0.68

Activations Density 0.010%

references to new or recent developments, such as updates or releases

latest news and updates

No Comments

No Known Activations