INDEX

Explanations

terms related to institutional entities and their affiliations

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

institutional affiliations

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

 mijne

-0.43

 księ

-0.43

 puncak

-0.43

 smlou

-0.37

 Verstär

-0.37

 Wikiseite

-0.37

 stále

-0.36

 stř

-0.36

wyg

-0.36

 pocz

-0.35

POSITIVE LOGITS

Personendaten

0.75

rungsseite

0.73

ScopeManager

0.71

 ſind

0.66

MigrationBuilder

0.63

 comuniques

0.63

 transférez

0.62

 Chwiliwch

0.62

 lenker

0.61

Tikang

0.61

Activations Density 0.001%

terms related to institutional entities and their affiliations

institutional affiliations

No Comments

No Known Activations