INDEX

Explanations

references to specific sections, data types, or elements related to programming and documentation

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

citations and uncommon words

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

 Könige

-0.39

UrlResolution

-0.34

 Belakang

-0.32

InstrumentedTest

-0.31

 Blätter

-0.31

 najbol

-0.30

 investigador

-0.30

 Contactez

-0.30

 виправивши

-0.30

 Decke

-0.30

POSITIVE LOGITS

 autorytatywna

0.77

 queſto

0.76

<unused41>

0.75

[@BOS@]

0.75

<unused68>

0.75

<unused16>

0.75

<unused51>

0.75

<unused47>

0.75

<unused8>

0.75

<unused14>

0.75

Activations Density 0.410%

references to specific sections, data types, or elements related to programming and documentation

citations and uncommon words

No Comments

No Known Activations