INDEX

Explanations

references to various types and categories within different contexts

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

ref-type attributes

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

 Paglinawan

-0.46

anglès

-0.45

 Wikiseite

-0.44

 wikipagina

-0.44

DataSnapshot

-0.41

Rüyada

-0.41

SharedDtor

-0.40

bkz

-0.40

锈钢

-0.39

 समीक्षाओं

-0.38

POSITIVE LOGITS

epam

0.45

eventType

0.40

 gynhyrchwyd

0.40

ouv

0.39

 distanci

0.39

TagHelpers

0.39

|}{}

0.38

Identyfik

0.38

ambién

0.36

انيف

0.36

Activations Density 0.277%

references to various types and categories within different contexts

ref-type attributes

No Comments

No Known Activations