INDEX

Explanations

repetitions of the word 'same'

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

*=-

-0.71

 Provided

-0.67

phasis

-0.65

Äĩ

-0.61

 Leilan

-0.60

ule

-0.60

Bei

-0.60

olate

-0.59

meet

-0.59

 Recomm

-0.58

POSITIVE LOGITS

 exact

1.03

 sized

0.97

 vein

0.96

 thing

0.95

 kinds

0.92

 amount

0.88

iating

0.88

 kind

0.83

 fate

0.83

 proportions

0.81

Activations Density 2.765%

repetitions of the word 'same'

No Comments

No Known Activations