INDEX

Explanations

code-related syntax and operations involving data manipulation

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

parsing and string manipulation

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

AddTagHelper

-0.64

 betweenstory

-0.61

الحياه

-0.59

transQ

-0.55

 IndexPath

-0.54

hasMoreElements

-0.52

upné

-0.51

 kasarigan

-0.51

iettivo

-0.51

 giustizia

-0.50

POSITIVE LOGITS

Externé

0.32

 nahilalakip

0.31

ORPG

0.29

mainly

0.28

 насељу

0.27

IBOutlet

0.26

yue

0.26

 tapis

0.26

 hydra

0.25

inee

0.25

Activations Density 0.014%

code-related syntax and operations involving data manipulation

parsing and string manipulation

No Comments

No Known Activations