INDEX

Explanations

terms and concepts related to statistical or mathematical modeling

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

programmer defined identifiers

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

impresa

-0.44

élimin

-0.43

Nw

-0.40

 OnInit

-0.37

Mic

-0.35

Cle

-0.35

ükemmel

-0.35

johtaja

-0.34

investissement

-0.34

 увиде

-0.33

POSITIVE LOGITS

fastjson

0.62

 astore

0.62

dataclass

0.60

AccessFile

0.60

 getItemId

0.60

PageIndex

0.59

 cade

0.58

tableFuture

0.57

 Dama

0.57

TabIndex

0.56

Activations Density 0.307%

terms and concepts related to statistical or mathematical modeling

programmer defined identifiers

No Comments

No Known Activations