INDEX

Explanations

references to extensive or detailed information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

ReusableCell

-0.56

dms

-0.56

 Jimmie

-0.55

 kingdom

-0.54

خاذ

-0.53

 Chord

-0.51

 Caen

-0.51

 isInitialized

-0.50

 Kingdom

-0.50

 judgments

-0.50

POSITIVE LOGITS

extensive

1.77

Extensive

1.73

 Extensive

1.72

 extensive

1.68

 extensively

1.38

considerable

1.16

Massive

1.15

 extenso

1.13

 Substantial

1.09

 umfang

1.08

Activations Density 0.127%

references to extensive or detailed information

No Comments

No Known Activations