INDEX

Explanations

terms related to enhancements or improvements in processes, designs, or technologies

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

enhanced/modified nouns

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

 coisa

-0.43

<bos>

-0.41

 يتيمه

-0.38

isFile

-0.36

ต่อ

-0.36

 immediately

-0.36

idUser

-0.36

 déplacer

-0.36

 vient

-0.35

gs

-0.35

POSITIVE LOGITS

 modified

0.73

hanced

0.71

 الرياضيه

0.70

modified

0.69

Modified

0.69

 revised

0.69

 Modified

0.67

 abridged

0.66

 Expanded

0.65

 Enhanced

0.65

Activations Density 0.039%

terms related to enhancements or improvements in processes, designs, or technologies

enhanced/modified nouns

No Comments

No Known Activations