INDEX

Explanations

references to specific dates and locations

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

onomy

-0.65

redo

-0.64

 Silence

-0.62

hra

-0.61

 silence

-0.59

gif

-0.59

zed

-0.58

sed

-0.58

peat

-0.58

rieg

-0.57

POSITIVE LOGITS

FINE

0.80

 tourist

0.71

 Canal

0.69

 Liberties

0.68

quartered

0.66

orate

0.66

ember

0.64

 Railway

0.64

 rainy

0.64

ukong

0.63

Activations Density 2.721%

references to specific dates and locations

No Comments

No Known Activations