INDEX

Explanations

references to software tools and code generation

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

automatic generation and tools

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



-0.46

참고

-0.39

Traducción

-0.35

erçe

-0.35

 Capacitor

-0.34

nack

-0.32

atario

-0.31

vada

-0.31

 personalità

-0.31

 încă

-0.31

POSITIVE LOGITS

contentLoaded

0.66

 CURIAM

0.62

RepeatedField

0.58

 automat

0.54

automatic

0.52

 automate

0.52

 automático

0.51

doGet

0.51

automated

0.49

 automatic

0.49

Activations Density 0.106%

references to software tools and code generation

automatic generation and tools

No Comments

No Known Activations