INDEX

Explanations

programming-related keywords and constructs

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

import tracim_backend

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

 للاسماء

-0.71

 незавершена

-0.68

UnusedPrivate

-0.65

Personensuche

-0.64

 ويكيميديا

-0.62

KURZBESCHREIBUNG

-0.62

invokeLater

-0.61

GraphicsUnit

-0.60

PreferredItem

-0.58

enumii

-0.58

POSITIVE LOGITS

 duração

0.44

iempos

0.44

 coroa

0.43

 kecelakaan

0.43

 score

0.41

 araña

0.40

autonomie

0.40

 χρό

0.40

 felices

0.40

 résoudre

0.40

Activations Density 0.018%

programming-related keywords and constructs

import tracim_backend

No Comments

No Known Activations