INDEX

Explanations

numerical values and their associations with ranked categories or scores

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

scale mapping values

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

 Budaya

-0.30

 macetas

-0.30

 igjen

-0.30

ř

-0.29

 Wellen

-0.29

 samband

-0.29

 inauguración

-0.28

 wraz

-0.28

 tamaños

-0.28

 infraestructura

-0.27

POSITIVE LOGITS

 للمعارف

0.68

0.65

 CreateTagHelper

0.61

 AssemblyTitle

0.60

 snippetHide

0.59

✢

0.59

bitField

0.59

GOTREF

0.58

ThroughAttribute

0.58

tagHelperRunner

0.57

Activations Density 0.080%

numerical values and their associations with ranked categories or scores

scale mapping values

No Comments

No Known Activations