INDEX

Explanations

punctuation and symbols used in quoted speech and narrative

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

future tense words

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

rollo

-0.49

ophylla

-0.49

fjspx

-0.46

libft

-0.42

OGND

-0.42

Seit

-0.42

 الرياضيه

-0.41

цездатний

-0.41

 belonged

-0.40

MONT

-0.40

POSITIVE LOGITS

 budou

0.66

 budú

0.64

 serão

0.63

 nantinya

0.60

 prochaines

0.58

 lesz

0.57

Será

0.57

 serán

0.57

 saranno

0.57

 künftig

0.57

Activations Density 0.029%

punctuation and symbols used in quoted speech and narrative

future tense words

No Comments

No Known Activations