INDEX

Explanations

punctuation marks and programming syntax elements

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

Tech Blocks

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

хьтан

-0.84

 Италијани

-0.76

 <<<<<<<<<<<<<<

-0.71

 EconPapers

-0.70

 unknownFields

-0.69

BufferException

-0.64

tonsoft

-0.63

RuleContext

-0.62

 typelib

-0.62

⟬

-0.60

POSITIVE LOGITS

 săn

0.35

 automática

0.29

 separado

0.28

 deshalb

0.28

گانه

0.27

 razie

0.27

เดิม

0.26

 piłkarz

0.25

 equivalente

0.25

 teda

0.25

Activations Density 0.097%

punctuation marks and programming syntax elements

Tech Blocks

No Comments

No Known Activations