INDEX

Explanations

punctuation marks and symbols

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

apache commons math

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

 CreateTagHelper

-0.76

exitRule

-0.69

tagext

-0.66

новниш

-0.66

 ujednoznacz

-0.62

oredCriteria

-0.59

parsedMessage

-0.59

transQ

-0.58

AndEndTag

-0.58

 Signalez

-0.57

POSITIVE LOGITS

 Mumbai

0.41

 vanilla

0.41

ppelin

0.40

 Plenum

0.39

 buffered

0.38

setminus

0.38

Vanilla

0.37

downloaded

0.37

shifted

0.36

 uncin

0.35

Activations Density 0.001%

punctuation marks and symbols

apache commons math

No Comments

No Known Activations