INDEX

Explanations

elements related to structure and categorization in data

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

technical code and documentation

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

DockStyle

-0.59

Bộ

-0.40

TRAILING

-0.40

ləş

-0.39

domada

-0.39

 AssemblyCompany

-0.39

Rujuakan

-0.38

localctx

-0.38

 вікісторінку

-0.38

 arşivlendi

-0.37

POSITIVE LOGITS

 informée

0.73

ValueStyle

0.46

ariConfig

0.43

astify

0.43

tagHelperRunner

0.42

ImageContext

0.41

Diweddarwch

0.40

Datuak

0.40

 puissiez

0.39

tvguidetime

0.39

Activations Density 0.097%

elements related to structure and categorization in data

technical code and documentation

No Comments

No Known Activations