INDEX

Explanations

terms related to legal pleas and convictions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.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

HtmlAttribute

-0.82

ISupport

-0.80

yoshi

-0.75

semi

-0.72

 messes

-0.70

Oss

-0.70

 mess

-0.68

box

-0.68

ेत्र

-0.67

bax

-0.67

POSITIVE LOGITS

Ho

1.18

Ho

1.16

ho

1.11

 hâte

0.86

HO

0.81

 Autre

0.81

 plead

0.80

guilty

0.79

Bead

0.79

 pleading

0.79

Activations Density 0.137%

terms related to legal pleas and convictions

No Comments

No Known Activations