INDEX

Explanations

references to legal concepts and actions

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

actions, deeds, and crimes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

DockStyle

-0.54

 GenerationType

-0.52

AccessorTable

-0.49

inSlope

-0.45

KEYCODE

-0.44

GenerationType

-0.44

CharSequence

-0.43

ReusableCell

-0.43

 AppCompatTheme

-0.43

IconModule

-0.43

POSITIVE LOGITS

 deeds

0.60

 crime

0.60

 crimes

0.59

 acts

0.59

 acte

0.52

crime

0.51

 deed

0.50

 performed

0.50

 actes

0.50

Crime

0.49

Activations Density 0.259%

references to legal concepts and actions

actions, deeds, and crimes

No Known Activations