INDEX

Explanations

key numeric values and decisions

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

sentence enders

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

 nahilalakip

-0.57

tagHelperRunner

-0.53

 Wicidata

-0.53

contentLoaded

-0.52

 CreateTagHelper

-0.52

 <=",

-0.51

 ویکی‌پدی

-0.50

хьтан

-0.50

ItemBackground

-0.49

 kasarigan

-0.46

POSITIVE LOGITS

 שוליים

0.40

 unters

0.37

usermodel

0.36

agg

0.34

SAP

0.34

UC

0.34

甲

0.34

ge

0.34

bua

0.33

偏偏

0.33

Activations Density 0.197%

key numeric values and decisions

sentence enders

No Comments

No Known Activations