INDEX

Explanations

legal and regulatory terminology related to claims and defenses

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

relationships between 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

and

-0.38

 assistência

-0.32

 anzu

-0.31

antikan

-0.31

 Verkehrs

-0.30

isinstance

-0.30

ิลป

-0.29

oreille

-0.29

 pouvoirs

-0.29

 terceiro

-0.29

POSITIVE LOGITS

 Comprometido

0.73

 Monfieur

0.73

 autorytatywna

0.72

Efq

0.68

ьаж

0.68

 ſta

0.66

ſelves

0.65

aarrggbb

0.65

 myſelf

0.64

SequentialGroup

0.64

Activations Density 0.080%

legal and regulatory terminology related to claims and defenses

relationships between words

No Comments

No Known Activations