INDEX

Explanations

terms related to finishing and completion, particularly in a creative or product context

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

finishing touches and products

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

Negative Logits

ronym

-0.45

 isNameExpr

-0.43

ystick

-0.43

usammen

-0.42

 presa

-0.42

InsertCommand

-0.41

 SIMBAD

-0.41

libatkan

-0.40

 Universitaria

-0.40

 unrelated

-0.39

POSITIVE LOGITS

 finishing

1.34

 Finishing

1.30

finishing

1.25

Finishing

1.23

 finish

1.16

finish

1.14

 Finish

1.08

 FINISH

1.05

Finish

1.02

FINISH

1.00

Activations Density 0.011%

terms related to finishing and completion, particularly in a creative or product context

finishing touches and products

No Known Activations