INDEX

Explanations

patterns involving repeated phrases or references to external elements, particularly focusing on articles and conjunctions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 13-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_13/width_65k/average_l0_75

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.13.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

 Chriftian

-0.57

 पत्र

-0.55

 gegeven

-0.55

ある

-0.54

 nød

-0.54

transQ

-0.53

estern

-0.53

Akt

-0.52

 medesimo

-0.52

 frattempo

-0.52

POSITIVE LOGITS

how

0.69

 better

0.69

 more

0.66

quê

0.65

 Exactos

0.65

älv

0.64

>::

0.62

berapa

0.62

 مرئيه

0.61

DockStyle

0.61

Activations Density 0.007%

patterns involving repeated phrases or references to external elements, particularly focusing on articles and conjunctions

No Comments

No Known Activations