INDEX

Explanations

instructions and actions related to navigating and utilizing features in an interface

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

on buttons and switches

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

UrlResolution

-0.88

AddTagHelper

-0.71

 navideño

-0.64

MessageOf

-0.62

 HttpNotFound

-0.61

EndGlobalSection

-0.61

 BoxFit

-0.61

 вікісторінку

-0.60

verifyException

-0.60

kegaard

-0.60

POSITIVE LOGITS

 hydrate

0.32

emo

0.31

gra

0.29

 arra

0.29

fl

0.29

 flip

0.28

哋

0.27

 flor

0.27

altham

0.27

 honor

0.26

Activations Density 0.030%

instructions and actions related to navigating and utilizing features in an interface

on buttons and switches

No Comments

No Known Activations