INDEX

Explanations

information related to health and medical procedures

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

wom

-0.61

 stre

-0.59

Bos

-0.56

"{

-0.56

 builder

-0.56

 arms

-0.54

PTS

-0.54

 mornings

-0.51

 abuser

-0.51

 âĵĺ

-0.50

POSITIVE LOGITS

1.51

1.50

1.35

1.34

1.33

1.31

1.30

Activations Density 2.046%

information related to health and medical procedures

No Comments

No Known Activations