INDEX

Explanations

interactions involving characters of authority and their responses or actions

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

subject speaking or expressing emotion

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

 Autorizaciones

-0.59

 EconPapers

-0.57

الحياه

-0.51

transQ

-0.51

 onCancelled

-0.50

قایناق‌لار

-0.50

 kasarigan

-0.50

 Autorisations

-0.50

Hochspringen

-0.49

 صوتيه

-0.49

POSITIVE LOGITS

 grinned

0.47

 chuckled

0.44

 blushed

0.43

 giggled

0.42

 smiled

0.41

 glared

0.39

 smirked

0.38

 sighed

0.38

 excitedly

0.37

 softly

0.37

Activations Density 0.012%

interactions involving characters of authority and their responses or actions

subject speaking or expressing emotion

No Comments

No Known Activations