INDEX

Explanations

web-related terms and concepts.

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hug

-0.84

 affili

-0.76

 pudding

-0.74

 advis

-0.73

 extinct

-0.73

 temperament

-0.73

 idle

-0.73

 renegoti

-0.72

 purse

-0.72

 defe

-0.72

POSITIVE LOGITS

 However

1.62

 Initially

1.55

 Previously

1.55

 Although

1.55

 They

1.53

 Though

1.53

Its

1.52

 Unfortunately

1.51

 While

1.50

 Each

1.50

Activations Density 3.602%

web-related terms and concepts.

No Comments

No Known Activations