INDEX

Explanations

references to political figures and statements on various topics

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

 upstream

-0.87

 downstream

-0.85

 extinct

-0.81

 waterfall

-0.78

 underground

-0.77

 harvest

-0.77

 commissions

-0.76

dra

-0.76

 orphan

-0.72

 probability

-0.72

POSITIVE LOGITS

Asked

1.80

Speaking

1.80

Refer

1.63

Calling

1.61

Talking

1.47

His

1.46

SPONSORED

1.46

During

1.43

Discuss

1.39

Writing

1.38

Activations Density 0.386%

references to political figures and statements on various topics

No Comments

No Known Activations