INDEX

Explanations

dates in the format of month followed by day with a high activation for January 9th

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

helle

-0.83

atics

-0.79

atic

-0.76

APH

-0.73

atis

-0.71

ribution

-0.71

hematic

-0.69

onse

-0.68

anced

-0.68

ropolitan

-0.67

POSITIVE LOGITS

nect

1.05

st

1.03

esville

1.03

vier

0.98

itors

0.88

itor

0.85

iotic

0.81

eker

0.78

ality

0.75

FK

0.74

Activations Density 1.292%

dates in the format of month followed by day with a high activation for January 9th

No Comments

No Known Activations