INDEX

Explanations

elements related to web content and URLs

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

title or blog post

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

Negative Logits

IntoConstraints

-0.42

GRANTED

-0.40

oredCriteria

-0.39

 jspb

-0.38

Jeografia

-0.37

webElementXpaths

-0.37

 المعيارى

-0.36

RuleContext

-0.35

braith

-0.35

henko

-0.35

POSITIVE LOGITS

 betweenstory

0.49

Fordítás

0.49

 Blog

0.45

发表于

0.45

Blog

0.45

 blog

0.44

(“

0.43

 lẽ

0.43

 titled

0.42

inerja

0.42

Activations Density 0.007%

elements related to web content and URLs

title or blog post

No Known Activations