INDEX

Explanations

numeric values, particularly quantities or counts

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

numbers followed by items

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

providedIn

-0.44

 OnInit

-0.40

 Wicidata

-0.38

Demografia

-0.38

aktur

-0.38

 enjeux

-0.36

ContentLoaded

-0.35

 речь

-0.35

 avantages

-0.35

POSITIVE LOGITS

AndEndTag

0.61

 ſeveral

0.60

 different

0.60

 dozen

0.58

ſelves

0.56

lapsingToolbar

0.56

Erreferent

0.54

duate

0.54

dozen

0.54

 eight

0.53

Activations Density 0.154%

numeric values, particularly quantities or counts

numbers followed by items

No Comments

No Known Activations