INDEX

Explanations

database-related terms and functions, as well as medical centers and their locations

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

 metic

-0.85

 acknow

-0.80

 estab

-0.78

 gelatin

-0.75

 polic

-0.74

 respected

-0.73

 Papers

-0.72

 belie

-0.72

 enriched

-0.72

 illustrated

-0.71

POSITIVE LOGITS

split

1.44

double

1.43

move

1.40

short

1.39

single

1.39

lower

1.39

requ

1.38

death

1.38

two

1.37

commun

1.37

Activations Density 1.878%

database-related terms and functions, as well as medical centers and their locations

No Comments

No Known Activations