INDEX

Explanations

references to scientific papers or articles, typically in a research context

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

** followed by digits

np_acts-logits-general · gemini-2.0-flash

numbers following years and journal references

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

ViewFeatures

-0.74

########.

-0.68

[]:

-0.66

 هيا

-0.57

はじめに

-0.55

 useDispatch

-0.54

Jeografia

-0.54

ThroughAttribute

-0.53

GenerationType

-0.53

toHaveBeenCalled

-0.51

POSITIVE LOGITS

 corners

0.59

 Senat

0.53

papi

0.53

 SHORT

0.52

 Short

0.52

 Corners

0.51

 noDo

0.50

titleMargin

0.49

 dagegen

0.49

Short

0.48

Activations Density 0.140%

references to scientific papers or articles, typically in a research context

** followed by digits

numbers following years and journal references

No Comments

No Known Activations