INDEX

Explanations

comparative phrases or expressions related to increase or decrease

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

more comparative

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

héro

-0.48

Nome

-0.46

kø

-0.45

 rövid

-0.43

jegy

-0.42

 enfans

-0.42

 DPRD

-0.41

wicklung

-0.41

 HasFactory

-0.41

 rękaw

-0.40

POSITIVE LOGITS

 more

0.89

 MORE

0.81

 más

0.81

 Lebih

0.75

 More

0.74

 più

0.73

 lebih

0.73

More

0.72

more

0.72

 Più

0.69

Activations Density 0.005%

comparative phrases or expressions related to increase or decrease

more comparative

No Comments

No Known Activations