INDEX

Explanations

sentences that contain key informational elements or clarifications

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

features, developments

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

 poffe

-0.46

culable

-0.41

izoen

-0.38

ſelf

-0.38

ſte

-0.36

acebook

-0.35

Definición

-0.35

 Italijanski

-0.35

cancellation

-0.35

 Appell

-0.35

POSITIVE LOGITS

UserScript

0.64

featureID

0.52

SequentialGroup

0.52

awakeFromNib

0.49

ftagPool

0.47

expandindo

0.46

 interesting

0.45

vrigt

0.44

flashdata

0.44

 kaarangay

0.42

Activations Density 0.084%

sentences that contain key informational elements or clarifications

features, developments

No Comments

No Known Activations