INDEX

Explanations

information related to workplace safety violations and legal citations

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

Redditor

-0.79

Laughs

-0.78

 sucker

-0.77

 flick

-0.75

 dudes

-0.74

 gigs

-0.74

Yep

-0.73

 dude

-0.72

 veggies

-0.71

 booze

-0.70

POSITIVE LOGITS

 certific

0.95

 Advisory

0.91

 Administrative

0.89

 Procedures

0.88

 pursuant

0.86

 Applicant

0.84

 exemplary

0.83

 Statement

0.82

 Effective

0.82

 Judicial

0.81

Activations Density 0.980%

information related to workplace safety violations and legal citations

No Comments

No Known Activations