INDEX

Explanations

mentions of politics and government policies

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

»Ĵ

-0.81

 bang

-0.72

yssey

-0.72

ascus

-0.71

 ogre

-0.70

 sustainable

-0.70

 pudding

-0.68

 unbeat

-0.67

 dimensional

-0.66

è¦ļéĨĴ

-0.65

POSITIVE LOGITS

 However

1.55

But

1.44

 Though

1.40

 They

1.36

 Nevertheless

1.36

 Although

1.35

 Instead

1.33

 Nonetheless

1.32

 Some

1.31

 While

1.31

Activations Density 1.881%

mentions of politics and government policies

No Comments

No Known Activations