INDEX

Explanations

terms related to computer science and technical tools, specifically focused on databases and data analysis

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

 politic

-0.73

çīĪ

-0.68

 Journalism

-0.63

democracy

-0.61

 democrat

-0.60

 resurg

-0.60

 millenn

-0.60

 Luxem

-0.59

 renaissance

-0.59

 Canaver

-0.58

POSITIVE LOGITS

 predetermined

0.76

 desired

0.70

 cursor

0.70

\'

0.67

 unwanted

0.66

 automatically

0.66

tx

0.65

or

0.64

 stimuli

0.64

 contact

0.64

Activations Density 5.541%

terms related to computer science and technical tools, specifically focused on databases and data analysis

No Comments

No Known Activations