INDEX

Explanations

locations, specifically named bodies of water

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

cious

-0.78

aign

-0.72

ISTER

-0.71

places

-0.70

nings

-0.69

book

-0.69

hered

-0.68

ebook

-0.68

acting

-0.67

fully

-0.67

POSITIVE LOGITS

 Yards

1.04

 Valley

0.95

Bay

0.85

 Strait

0.83

 Basin

0.79

 Falls

0.77

dale

0.76

XT

0.74

burgh

0.74

 Highlands

0.73

Activations Density 0.026%

locations, specifically named bodies of water

No Comments

No Known Activations