INDEX

Explanations

tokens or commands that indicate formatting or style instructions in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 7-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_7/width_16k/average_l0_69

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.7.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

ません

-0.75

 Pfalz

-0.72

 atos

-0.71

Def

-0.70

Topology

-0.69

 Culver

-0.67

ashi

-0.64

 dage

-0.63

 Manhattan

-0.62

Vill

-0.62

POSITIVE LOGITS

{~

1.98

--}}

0.90

wisata

0.89

mallows

0.86

 Sampler

0.85

 IBOutlet

0.82

 GenerationType

0.80

הערות

0.80

 useCallback

0.79

 henvisninger

0.79

Activations Density 0.059%

tokens or commands that indicate formatting or style instructions in text

No Comments

No Known Activations