INDEX

Explanations

repeated numerical references, particularly percentages and statistical data

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

Multilingual strings

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

 increí

-0.90

 queſta

-0.88

StoryboardSegue

-0.86

 indígen

-0.84

MigrationBuilder

-0.83

 incrí

-0.82

 dezelve

-0.82

 propOrder

-0.81

 AssemblyCulture

-0.79



-0.78

POSITIVE LOGITS

Bis

0.29

↵

0.27

0.26

Còn

0.25

[*]

0.25

www

0.24

Super

0.23

Activations Density 0.000%

repeated numerical references, particularly percentages and statistical data

Multilingual strings

No Comments

No Known Activations