INDEX

Explanations

programming concepts and terminology, particularly related to task scheduling and handling errors

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

programming code snippets

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

ganggu

-0.38

 héroes

-0.32

 peindre

-0.31

 Friedhof

-0.31

 Ruhe

-0.31

 héros

-0.30

 Hoffnung

-0.30

"""

-0.30

POSITIVE LOGITS

twimg

1.41

 ब्रेकडाउन

1.28

 незавершена

1.24

uxxxx

1.20

 Мексичка

1.13

oredCriteria

1.12

AddTagHelper

1.12

styleType

1.10

 ModelExpression

1.10

 CreateTagHelper

1.09

Activations Density 0.002%

programming concepts and terminology, particularly related to task scheduling and handling errors

programming code snippets

No Comments

No Known Activations