INDEX

Explanations

elements related to mathematical or programming structures, particularly those indicating relationships or conditions

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

C++ code snippets

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

MLLoader

-0.66

 feroit

-0.63

awtextra

-0.62

TypedDataSet

-0.62

 pouvoit

-0.59

 ſont

-0.58

 auroit

-0.58

 ainfi

-0.58

 myſelf

-0.57

 queſta

-0.57

POSITIVE LOGITS

MemoryWarning

0.40

 henvisninger

0.34

SharedDtor

0.33

hyrchwyd

0.31

かせて

0.31



0.31

ylus

0.31

 baking

0.31

 inherent

0.30

unjung

0.29

Activations Density 0.025%

elements related to mathematical or programming structures, particularly those indicating relationships or conditions

C++ code snippets

No Comments

No Known Activations