INDEX

Explanations

references to pages and printed documents

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

specific page numbers

np_acts-logits-general · gemini-2.0-flash

page numbers

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

исленность

-0.36

thâu

-0.35

cember

-0.32

 descal

-0.31

 termin

-0.31

 dece

-0.30

发表于

-0.30

 Medicines

-0.30

sed

-0.30

一支

-0.30

POSITIVE LOGITS

 pages

0.88

 page

0.79

ChildScrollView

0.75

 sheet

0.74

 Page

0.72

pages

0.72

 sheets

0.71

page

0.71

Page

0.68

 Pages

0.66

Activations Density 0.519%

references to pages and printed documents

specific page numbers

page numbers

No Comments

No Known Activations