INDEX

Explanations

terms related to data, particularly personal information and its implications

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

ãĤ¢ãĥ«

-0.81

ITH

-0.75

ãĤ¨ãĥ«

-0.73

Tur

-0.72

STER

-0.70

ãĥ¤

-0.70

ãĥį

-0.69

Ô

-0.69

CHAR

-0.69

sonian

-0.69

POSITIVE LOGITS

hips

1.08

etting

1.00

ynthesis

1.00

etc

0.95

avers

0.91

hip

0.89

mith

0.87

hops

0.83

gal

0.80

 abound

0.80

Activations Density 0.276%

terms related to data, particularly personal information and its implications

No Comments

No Known Activations