INDEX

Explanations

descriptions of physical objects and their attributes, particularly related to fitting and construction

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 credential

-0.88

 waged

-0.88

 divest

-0.79

 unle

-0.78

 enshr

-0.77

 mobil

-0.77

 counterterrorism

-0.77

 enrol

-0.75

 repud

-0.74

 wage

-0.74

POSITIVE LOGITS

Anyway

1.69

 Anyway

1.67

 Honestly

1.36

 haha

1.35

:)

1.34

;)

1.28

 Personally

1.28

 Hopefully

1.27

:(

1.27

Honestly

1.27

Activations Density 4.441%

descriptions of physical objects and their attributes, particularly related to fitting and construction

No Comments

No Known Activations