INDEX

Explanations

numeric data and statistical findings related to experimental results

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

5 grouping smallest

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

Numerology

-0.34

といけない

-0.33

dz

-0.32

Requirements

-0.32

 dég

-0.31

Qual

-0.30

ernoo

-0.30

手が

-0.30

 insu

-0.30

Dear

-0.29

POSITIVE LOGITS

 للاسماء

0.57

HtmlAttribute

0.55

 healthiest

0.54

 préférence

0.52

disposing

0.52

 Normdatei

0.51

 незавершена

0.50

 ModelAndView

0.50

bootstrapcdn

0.50

 fairest

0.49

Activations Density 0.471%

numeric data and statistical findings related to experimental results

5 grouping smallest

No Comments

No Known Activations