INDEX

Explanations

phrases related to instructions or purposes for design in products or systems

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-2b-pt-res/layer_21/width_16k/average_l0_70

Prompts (Dashboard)

36,864 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

abancı

-0.62

टक

-0.57

rê

-0.51

abestanden

-0.51

machung

-0.51

jsonwebtoken

-0.51

 Somebody

-0.50

SearchParams

-0.49

 groupBox

-0.49

DebuggerStep

-0.48

POSITIVE LOGITS

 ensure

1.00

 maximize

0.81

 provide

0.81

 ensures

0.77

 guarantee

0.76

ensure

0.74

 minimize

0.73

 ensuring

0.72

 Ensure

0.72

 insure

0.69

Activations Density 0.332%

phrases related to instructions or purposes for design in products or systems

No Comments

No Known Activations