INDEX

Explanations

prepositions followed by geopolitical location mentions

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

neys

-0.62

 reacts

-0.62

AMA

-0.61

APD

-0.60

 Flavoring

-0.59

 batted

-0.58

 succeed

-0.56

 Doodle

-0.56

 greets

-0.56

 interacts

-0.55

POSITIVE LOGITS

lined

0.94

escap

0.91

versible

0.87

uble

0.85

avering

0.84

bred

0.83

ked

0.82

stated

0.81

putable

0.78

lain

0.78

Activations Density 1.231%

prepositions followed by geopolitical location mentions

No Comments

No Known Activations