INDEX

Explanations

physical locations and addresses

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 carbohyd

-0.80

behind

-0.76

 sympathetic

-0.69

 mathemat

-0.68

PDATE

-0.67

 variance

-0.67

meric

-0.65

 situational

-0.64

 millenn

-0.64

 conduc

-0.63

POSITIVE LOGITS

 County

1.29

 Creek

1.18

 Abbey

1.17

 Valley

1.15

 Islands

1.14

 Township

1.14

 Manor

1.13

 Island

1.12

 Farms

1.11

 Falls

1.10

Activations Density 2.834%

physical locations and addresses

No Comments

No Known Activations