INDEX

Explanations

terms related to legal and correctional processes

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

poll

-0.70

anto

-0.70

fters

-0.66

 Continent

-0.60

 gener

-0.60

metics

-0.60

bags

-0.59

 audible

-0.59

zi

-0.58

bag

-0.58

POSITIVE LOGITS

ilitating

1.39

ilitation

1.30

 probation

1.04

 rehabilitation

0.98

 rehabilit

0.90

 supervision

0.86

 parole

0.86

 Rehab

0.81

spective

0.80

 rehab

0.78

Activations Density 0.042%

terms related to legal and correctional processes

No Comments

No Known Activations