INDEX

Explanations

patterns of numbers and values commonly associated with structured data or programming elements

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

HTML code snippets

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

 ſind

-1.21

 queſta

-1.17

iſen

-1.16

parsedMessage

-1.16

majánló

-1.15

niſſe

-1.13

ésultats

-1.13

ロウィン

-1.09

ſicht

-1.09

ſſung

-1.08

POSITIVE LOGITS

0.59

0.57

0.52

[toxicity=0]

0.50

0.46

0.45

↵

0.45

Activations Density 0.041%

patterns of numbers and values commonly associated with structured data or programming elements

HTML code snippets

No Comments

No Known Activations