INDEX

Explanations

references to previous discussions, articles, and content within a text

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

document references

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

 satisfacción

-0.40

 acrylique

-0.38

 Universitario

-0.38

 conducción

-0.37

 comodidad

-0.37

 felicidad

-0.37

alingen

-0.36

 naturaleza

-0.36

 supervivencia

-0.35

 Sloveniji

-0.35

POSITIVE LOGITS

RTEE

0.85

 виправивши

0.69

verwijspagina

0.66

Життєпис

0.62

 يتيمه

0.58

 disambiguazione

0.57

:✨

0.56

OGND

0.55

 ProtoMessage

0.54

Diweddarwch

0.54

Activations Density 0.278%

references to previous discussions, articles, and content within a text

document references

No Comments

No Known Activations