INDEX

Explanations

critical legal terms and phrases in court-related contexts

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

when making choices/decisions

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

 nonUne

-0.57

 defaultstate

-0.54

ArrowToggle

-0.54

 zwiſchen

-0.54

 deſſen

-0.53

 برانيه

-0.52

extAlignment

-0.52

writeField

-0.52

'{@

-0.52

MLLoader

-0.51

POSITIVE LOGITS

 always

0.45

 Dabei

0.45

 choices

0.40

 must

0.39

 attention

0.38

 often

0.37

 involves

0.37

 waarbij

0.36

 consideration

0.36

Dabei

0.35

Activations Density 0.055%

critical legal terms and phrases in court-related contexts

when making choices/decisions

No Comments

No Known Activations