INDEX

Explanations

scientific and technical language related to biological and medical research

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

 tack

-0.82

 silly

-0.80

 dime

-0.77

lier

-0.75

 grip

-0.74

plet

-0.74

hug

-0.74

 quir

-0.73

 instinct

-0.73

 ugly

-0.72

POSITIVE LOGITS

 Additionally

1.94

 Furthermore

1.83

 However

1.75

 Therefore

1.73

 Moreover

1.72

 Consequently

1.68

 Accordingly

1.65

 Thus

1.56

 Similarly

1.56

 Likewise

1.51

Activations Density 1.694%

scientific and technical language related to biological and medical research

No Comments

No Known Activations