INDEX

Explanations

proper nouns, specifically names and entities related to people and places

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

names followed by context

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-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

 Esposito

-0.49

happen

-0.48

 Jennings

-0.47

Lors

-0.47

 causes

-0.47

 tries

-0.47

 occurrence

-0.45

 rounds

-0.45

 Jensen

-0.45

 Madsen

-0.44

POSITIVE LOGITS

 <<<<<<<<<<<<<<

0.62

 EconPapers

0.60

*/].

0.59

</tfoot>

0.58

 inoxydable

0.57

DotNetBar

0.57

󠁮

0.54

новниш

0.54

 épaules

0.54

 cheminée

0.52

Activations Density 0.443%

proper nouns, specifically names and entities related to people and places

names followed by context

No Comments

No Known Activations