INDEX

Explanations

mentions of time periods, specifically years and weeks

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

earlier this year, week, or month

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 habet

-0.30

 skär

-0.28

 âmes

-0.26

 plazos

-0.25

 enfans

-0.25

 reda

-0.25

 étoit

-0.23

 peuples

-0.22

 proyec

-0.22

 chré

-0.21

POSITIVE LOGITS

 nakalista

0.82

المناصب

0.81



0.78

 '\\;'

0.73

uxxxx

0.73

tagHelperRunner

0.73

ArgumentParser

0.72

𓆸

0.71

࿊

0.71

 فريبيس

0.71

Activations Density 0.003%

mentions of time periods, specifically years and weeks

earlier this year, week, or month

No Known Activations