INDEX

Explanations

words related to leadership and leading actions or initiatives

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

 loopholes

-0.69

afort

-0.67

 privilege

-0.66

entric

-0.64

perture

-0.62

aves

-0.61

ukong

-0.60

pmwiki

-0.60

dam

-0.60

 grace

-0.58

POSITIVE LOGITS

gers

0.89

hunt

0.87

ging

0.81

better

0.80

owitz

0.78

ership

0.78

leaders

0.74

TAIN

0.72

tyard

0.72

leading

0.71

Activations Density 4.257%

words related to leadership and leading actions or initiatives

No Comments

No Known Activations