INDEX

Explanations

references to locations or conditions indicated by the word "where."

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

"where"

np_acts-logits-general · gemini-2.0-flash

introduces context after "where"

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_16k/average_l0_88

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.9.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

DPS

-0.47

onOptions

-0.44

DPS

-0.43

 NSMutable

-0.43

Offensive

-0.43

DEMO

-0.42

Adu

-0.42

 Ekiti

-0.42

Puff

-0.42

 Puff

-0.41

POSITIVE LOGITS

 where

1.77

where

1.70

Where

1.55

 Where

1.54

 WHERE

1.44

WHERE

1.42

 donde

1.29

Dónde

1.23

où

1.23

 dónde

1.22

Activations Density 0.165%

references to locations or conditions indicated by the word "where."

"where"

introduces context after "where"

No Comments

No Known Activations