INDEX

Explanations

patterns related to numerical information, particularly involving the number four

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

The number "4" followed by various units or descriptors

np_acts-logits-general · gemini-2.0-flash

4 followed by digits or words

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

TypedDataSet

-0.59

 saites

-0.58

LabelTagHelper

-0.57

helf

-0.57

oneofs

-0.56

$]$

-0.55

 }}"></

-0.54

Demografia

-0.53

AndroidJUnit

-0.53

}</

-0.51

POSITIVE LOGITS

 legged

0.69

legged

0.60

½

0.55

yrs

0.55

️⃣

0.53

 reasons

0.53

 different

0.51

 main

0.51

 consecutive

0.50

 dimensional

0.50

Activations Density 0.589%

patterns related to numerical information, particularly involving the number four

The number "4" followed by various units or descriptors

4 followed by digits or words

No Comments

No Known Activations