INDEX

Explanations

terms related to inmate communication rules and privileges

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

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

 autorytatywna

-0.80

zyg

-0.58

fycat

-0.54

atase

-0.52

Попис

-0.51

 beginnetje

-0.49

prüche

-0.49

andidaten

-0.49

Controllo

-0.49

hashtags

-0.47

POSITIVE LOGITS

 letter

2.35

 letters

2.21

letter

1.95

 Letters

1.94

Letter

1.92

 Letter

1.92

 LETTER

1.88

letters

1.87

Letters

1.81

 mail

1.74

Activations Density 0.450%

terms related to inmate communication rules and privileges

No Comments

No Known Activations