INDEX

Explanations

terms related to invention and milestones in history

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

first or early instance

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

AddTagHelper

-0.65

 referenties

-0.60

MemoryWarning

-0.57

hyrchwyd

-0.56

kuuta

-0.52

RefNanny

-0.52

hésite

-0.51

RenderAtEndOf

-0.49

jectives

-0.49

égard

-0.48

POSITIVE LOGITS

 earliest

0.53

最早

0.48

 first

0.40

 мәкалә

0.39

最初に

0.39

 pertama

0.38

 первая

0.37

 ersten

0.37

最初

0.37

 primeira

0.36

Activations Density 0.047%

terms related to invention and milestones in history

first or early instance

No Comments

No Known Activations