INDEX

Explanations

legal and court-related terminology

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

Followed by a comma

np_acts-logits-general · gemini-2.0-flash

via

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 unto

-0.93

ParallelGroup

-0.83

 kepada

-0.80

 vnto

-0.77

ագրություններ

-0.74

给

-0.70

StringTo

-0.70

EndInit

-0.70

向

-0.67

 Unto

-0.67

POSITIVE LOGITS

via

0.62

يكب

0.47

Via

0.45

pygame

0.45

via

0.45

 vía

0.44

 través

0.43

keV

0.41

Via

0.41

 tramite

0.39

Activations Density 0.337%

legal and court-related terminology

Followed by a comma

via

No Comments

No Known Activations