INDEX

Explanations

legal terminology and references to judicial processes

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

Tokens preceding words like "not," "grant," or "deny."

np_acts-logits-general · gemini-2.0-flash

granting or denying permissions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

escaping

-0.54

 escaping

-0.51

escape

-0.51

 Survi

-0.48

 undergoes

-0.48

escaped

-0.48

 escapes

-0.47

 попа

-0.47

 underwent

-0.46

Escape

-0.46

POSITIVE LOGITS

 forbid

0.87

 granting

0.87

 grant

0.81

 allow

0.78

oa̍t

0.78

 impose

0.77

 approve

0.74

 awarding

0.72

 sanction

0.71

grant

0.70

Activations Density 0.890%

legal terminology and references to judicial processes

Tokens preceding words like "not," "grant," or "deny."

granting or denying permissions

No Comments

No Known Activations