INDEX

Explanations

structured conversation and interaction cues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 9-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_9/width_16k/average_l0_73

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

Not in Any Lists

No Comments

Negative Logits

 purpoſe

-0.64

 Shakspeare

-0.62

 Administrativna

-0.61

存于互联网档案馆

-0.60

 cauſe

-0.59

 Diſ

-0.59

 habet

-0.57

*/

-0.57

 themſelves

-0.57

 Jefus

-0.56

POSITIVE LOGITS

findpost

0.73

 发表于

0.72

Quote

0.70

RemoveField

0.62

Zitat

0.60

 للمعارف

0.56

SourceChecksum

0.55

 GenerationType

0.50

çoise

0.49

0.48

Activations Density 0.362%

structured conversation and interaction cues

No Comments

No Known Activations