INDEX

Explanations

terms related to apo-proteins and their variants

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

words like apology or apocalyptic

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

mtr

-0.44

 Marten

-0.43

hline

-0.41

Verwendung

-0.39

Redacción

-0.39



-0.38

</thead>

-0.38

makeConstraints

-0.38

‍♀️

-0.37

mtd

-0.37

POSITIVE LOGITS

Apo

0.97

Apo

0.85

apo

0.81

Apoll

0.81

apo

0.76

APO

0.73

 Apoll

0.70

 apologies

0.68

 Apost

0.67

 apology

0.67

Activations Density 0.015%

terms related to apo-proteins and their variants

words like apology or apocalyptic

No Comments

No Known Activations