INDEX

Explanations

actions related to writing and communication

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

write what comes next

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

numerusform

-0.55

 parezca

-0.48

 rajut

-0.48

 navideños

-0.46

 consecuencia

-0.44

 kompeti

-0.44

llavero

-0.43

vician

-0.43

 recomienda

-0.42

ulongan

-0.42

POSITIVE LOGITS

دانشنامهٔ

0.51

 arşivlendi

0.47

 nakalista

0.43

 PeEnEo

0.42

expandindo

0.41

tagHelperRunner

0.40

この記事

0.38

mybatisplus

0.38

AddTagHelper

0.38

AndroidJUnit

0.37

Activations Density 0.151%

actions related to writing and communication

write what comes next

No Comments

No Known Activations