INDEX

Explanations

occurrences of relational and defining phrases in discussions

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

functional results or criminal mischief

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

 Basket

-0.45

 المعيارى

-0.44

 Shel

-0.44

 hosts

-0.42

 strconv

-0.42

\|_{

-0.42

culata

-0.42

Teg

-0.41

Cad

-0.41

lein

-0.41

POSITIVE LOGITS

 Infór

0.56

 estatua

0.54

 TAMBÉM

0.52

 CURIAM

0.52

 plástica

0.52

MessageTagHelper

0.51

 cimetière

0.50

 AssemblyCulture

0.50

 Italij

0.49

 cerámica

0.49

Activations Density 0.004%

occurrences of relational and defining phrases in discussions

functional results or criminal mischief

No Comments

No Known Activations