INDEX

Explanations

terms related to experimental findings and indications of relationships or effects

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

suggests increased or impaired activity

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

PIXEL

-0.31

CppMethod

-0.30

 frère

-0.29

Referanser

-0.28

 vyst

-0.27

 suyo

-0.27

甚

-0.27

 hermanos

-0.27

reporte

-0.27

 débats

-0.27

POSITIVE LOGITS

 EconPapers

0.82

 CreateTagHelper

0.69

 Administrativna

0.63

rizona

0.59



0.54

queryInterface

0.53

AndEndTag

0.52

rrggbb

0.52

aarrggbb

0.52

vací

0.52

Activations Density 0.151%

terms related to experimental findings and indications of relationships or effects

suggests increased or impaired activity

No Comments

No Known Activations