INDEX

Explanations

descriptions of medical procedures and healthcare incidents

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

Trend

-0.73

anwhile

-0.64

 appropri

-0.64

HCR

-0.62

 behaviours

-0.59

 surrog

-0.59

 concess

-0.58

berus

-0.57

lements

-0.57

issance

-0.57

POSITIVE LOGITS

heartedly

1.09

dozen

0.87

 dozen

0.86

pipe

0.85

hearted

0.83

imet

0.81

wheel

0.78

moon

0.76

assed

0.76

wit

0.76

Activations Density 0.418%

descriptions of medical procedures and healthcare incidents

No Comments

No Known Activations