INDEX

Explanations

segments of text related to structured data or programming

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

hexadecimal and code structures

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

InjectAttribute

-0.59

 AspNetCore

-0.59

RegressionTest

-0.58

ChildScrollView

-0.57

 Мексичка

-0.53

原始内容存档

-0.52

 bezeichneter

-0.52

fromnode

-0.51

SourceChecksum

-0.50

Rhestr

-0.50

POSITIVE LOGITS

новништво

0.40

VolleyError

0.35

Tener

0.35

 earlier

0.33

WebMethod

0.32

Streets

0.32

 initially

0.31

 indic

0.31

verticalLayout

0.31

 STEEL

0.30

Activations Density 0.157%

segments of text related to structured data or programming

hexadecimal and code structures

No Comments

No Known Activations