INDEX

Explanations

technical and medical information

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

enei

-0.59

Registered

-0.59

Deal

-0.56

leneck

-0.56

 Golf

-0.54

iltr

-0.54

 tops

-0.54

Adult

-0.53

 Fighter

-0.52

 Males

-0.52

POSITIVE LOGITS

 others

0.94

 countless

0.89

ersen

0.82

romeda

0.80

 subsequent

0.80

isSpecialOrderable

0.76

 innumerable

0.74

 deserve

0.72

 untold

0.70

 elsewhere

0.69

Activations Density 0.100%

technical and medical information

No Comments

No Known Activations