INDEX

Explanations

references to memoirs and autobiographical content

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

 constitu

-0.82

ulhu

-0.70

coord

-0.70

OPA

-0.68

 jurisd

-0.65

 unintended

-0.63

 Reserved

-0.62

pots

-0.60

 retard

-0.60

yg

-0.60

POSITIVE LOGITS

 memoir

1.27

 autobiography

1.12

oir

0.92

lishing

0.91

 autobi

0.91

ogue

0.88

writer

0.87

writers

0.84

spective

0.82

 essay

0.81

Activations Density 0.012%

references to memoirs and autobiographical content

No Comments

No Known Activations