INDEX

Explanations

mentions of the word "Sens" with varying activation values

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Zimbabwe

-0.76

 Springer

-0.66

 Carnegie

-0.64

 Maker

-0.64

maker

-0.64

 Georgian

-0.62

Ala

-0.61

 Newton

-0.58

 Ministers

-0.58

 Gamb

-0.58

POSITIVE LOGITS

itivity

1.36

itives

1.28

itive

1.23

ibly

1.18

itized

1.15

atile

1.13

ibility

1.10

uous

1.08

itiveness

1.06

orial

1.05

Activations Density 0.013%

mentions of the word "Sens" with varying activation values

No Comments

No Known Activations