INDEX

Explanations

important named entities and their numerical or categorical significance

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

equations and variables

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

cipar

-0.39

di

-0.38

Hauptartikel

-0.38

zzleHttp

-0.37

後半

-0.37

 flatter

-0.36

 suckers

-0.36

 Beine

-0.36

 atom

-0.36

 buzz

-0.36

POSITIVE LOGITS

 kasarigan

0.80

tagHelperRunner

0.69

stdc

0.54

 Chwiliwch

0.52

 resourceCulture

0.51

SequentialGroup

0.50

 שוליים

0.48

❋

0.48

 autorytatywna

0.47

 queſta

0.46

Activations Density 0.001%

important named entities and their numerical or categorical significance

equations and variables

No Comments

No Known Activations