INDEX

Explanations

occurrences of the word "recent" and its variations, indicating a focus on the timeliness or newness of information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 propOrder

-0.68

AxisAlignment

-0.52

ConstraintMaker

-0.52

 pleaſure

-0.50

oredCriteria

-0.48

⟬

-0.46

GenerationType

-0.46

 ball

-0.45

illoma

-0.45

 armature

-0.44

POSITIVE LOGITS

recent

1.04

 RECENT

1.00

Recent

0.98

 recent

0.97

 Recent

0.94

recently

0.91

RECENT

0.90

 reciente

0.86

 recently

0.85

 récente

0.83

Activations Density 0.057%

occurrences of the word "recent" and its variations, indicating a focus on the timeliness or newness of information

No Comments

No Known Activations