INDEX

Explanations

numerical quantities related to statistics or percentages

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

ande

-0.74

complex

-0.73

imir

-0.69

sort

-0.69

rev

-0.67

gang

-0.66

 ramp

-0.66

heng

-0.64

Stage

-0.63

BIT

-0.63

POSITIVE LOGITS

 Americans

0.97

 respondents

0.89

 households

0.87

 earners

0.82

 adults

0.78

 Australians

0.77

 Britons

0.76

 American

0.74

 occurrences

0.72

 cases

0.71

Activations Density 0.051%

numerical quantities related to statistics or percentages

No Comments

No Known Activations