INDEX

Explanations

occurrences of the word "que" and fragments indicating quotation or code structure

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

technical discussions and code

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

tvguidetime

-1.09



-1.08

 AssemblyCompany

-1.03

出版年

-1.03

 queſta

-1.02

richTextPanel

-0.94

تقاوى

-0.94

iſen

-0.94

webElementXpaths

-0.93

 snippetHide

-0.93

POSITIVE LOGITS

0.48

[toxicity=0]

0.45

ing

0.43

0.40

ine

0.40

">(</

0.40

0.39

the

0.39

 числе

0.38

Activations Density 0.000%

occurrences of the word "que" and fragments indicating quotation or code structure

technical discussions and code

No Comments

No Known Activations