INDEX

Explanations

terms related to design and structured approaches in various contexts

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

specific noun phrases

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

Nas

-0.35

 fist

-0.33

 beating

-0.31

assertEquals

-0.30

 diputados

-0.30

anstal

-0.30

íncia

-0.30

を使用した

-0.29

ọng

-0.29

代わりに

-0.28

POSITIVE LOGITS

 Numerade

0.72

IUrlHelper

0.68

 AssemblyCulture

0.64

WithIOException

0.64

 PeEnEo

0.63

 autorytatywna

0.62

 propOrder

0.62

providedIn

0.62

省市镇

0.59

majánló

0.57

Activations Density 0.412%

terms related to design and structured approaches in various contexts

specific noun phrases

No Comments

No Known Activations