INDEX

Explanations

specific symbols or codes associated with data metrics or configurations

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

hexadecimal numbers and symbols

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

tagHelperRunner

-0.72

fjspx

-0.65

 Wikimédia

-0.65

aarrggbb

-0.65

 الحره

-0.60

bootstrapcdn

-0.60

 مرئيه

-0.59

PyExc

-0.56

tvguidetime

-0.56

transQ

-0.56

POSITIVE LOGITS

libatkan

0.36

 ungkapnya

0.34

umpul

0.34

Cuántos

0.34

 ucapnya

0.33

 synes

0.33

ByExample

0.32

 næste

0.32

 plis

0.32

 Rathaus

0.32

Activations Density 0.001%

specific symbols or codes associated with data metrics or configurations

hexadecimal numbers and symbols

No Comments

No Known Activations